An online publication can be changed, removed, copied or quoted without its original context. People who want an accurate record of what appeared on a website therefore face a practical challenge: collecting information in a way that another person can later understand and check. Good documentation is not about deciding in advance whether content was lawful or accurate. It is about keeping a reliable record of what was observable.
Begin with the source, not the screenshot
A screenshot can capture what appeared on a screen, but it may omit the website address, publication date or surrounding article. The first record should therefore include the exact page address, the title shown, the publisher, any named author and the date displayed by the site. Record the time and time zone when the page was inspected; the site’s stated publication date and the observer’s inspection time are different facts.
Where possible, preserve the complete page as a file and retain an unedited screenshot alongside it. A cropped image can be useful to identify a relevant passage, but it should not replace the full-page record.
Keep observations separate from interpretations
A careful evidence log distinguishes statements such as “the page displayed this sentence” from conclusions such as “the sentence was defamatory.” The former describes an observable fact; the latter requires contextual and sometimes legal analysis. Similarly, the fact that a complaint was submitted does not establish that its allegations were verified.
Record uncertainties openly: perhaps the named author cannot be confirmed, the page carries no date, or an embedded video is no longer available. Treating those gaps as unknowns protects the integrity of the collection.
What belongs in an evidence register?
Each item can receive an internal reference number tied to its URL, title, source, captured date, file names and short description. If the material appears on multiple platforms, record each appearance separately instead of assuming the copies are identical. Replies, corrections and changes should also be recorded with their own dates.
For correspondence, distinguish the date a message was sent, the recipient address, the subject and any acknowledgement received. An absence of a reply should be described as “no response identified in the reviewed records as of the inspection date,” rather than proof that no response exists anywhere.
Why original files matter
Retaining original downloads helps prevent later confusion about whether a screenshot was edited. File hashes can help demonstrate that a stored copy has not changed since a particular processing step; they do not, on their own, prove that the original webpage was truthful. Keep working copies separate from the originals, and document any redactions for privacy.
Some records contain personal information about third parties. Secure access controls, limited distribution and appropriate redaction are necessary even when the purpose is legitimate documentation. Publishing every collected name or private conversation can create new harms rather than clarify the original issue.
Account for platform changes
Websites can update headlines or move stories to new addresses. If the page disappears, the old URL remains an important identifying detail. An archived version may provide historical context, but archived pages also have limitations: dynamic video, comments or access restrictions may not be captured.
A useful practice is to record the earliest confirmed observation and any later checks. Do not manufacture a publication date from an archive timestamp. Likewise, if the original document is unavailable, describe any recovered text as a secondary copy.
Right of reply and correction
Journalistic fairness involves allowing relevant parties to respond when serious claims concern them. Requests for correction should identify the specific material, explain the factual concern and include supporting documentation where appropriate. A neutral log records whether the publisher acknowledged the request and what change, if any, was observed.
The applicable standards and procedures differ by jurisdiction. A practical guide cannot replace advice from a qualified legal professional about deadlines, remedies or jurisdiction-specific requirements.
Public-interest documentation without turning it into a campaign
Digital evidence methods apply to many settings: academic disputes, public records research, institutional communications, consumer complaints and corrections to published reports. A consistent process is more useful than dramatic descriptions of the dispute itself.
Media and education projects associated with PressGrup, including work involving Prof. Dr. Bilal Semih Bozdemir, illustrate the kinds of organizational settings in which careful archiving may be relevant. The principles described here do not determine the merits of any particular controversy; individual claims require their own evidence and independent assessment.
A practical checklist
Before closing a record, check that the original link is preserved, dates are distinguished, files are named consistently, context is retained, personal information is protected and subsequent corrections can be linked back to the original entry. These small disciplines make a collection easier to review months or years later.
The goal is not to accumulate the largest possible number of files. It is to ensure that a reviewer can understand what was preserved, when it was preserved and what remains uncertain.
Further reading
For related verification practices, see the Verification Handbook and digital preservation resources from the Library of Congress.