How can I tell whether an online source has direct knowledge or is reporting someone else’s account?

Start by separating two questions: “How does this source know?” and “Is the claim true?” The first is about access and attribution; the second needs its own evaluation. A person can describe what they say they saw, while a careful reporter may rely on a named witness. Neither connection alone settles whether every detail is correct.

For each item, write down the exact claim in plain words and the source’s stated basis for it. Look for phrases such as “I saw,” “according to [named person],” or “the outlet reports, citing [document or source].” Be cautious when a post gives no basis, or when an article merely says “reports say” without identifying them. Don’t upgrade “someone said this happened” into “this happened.”

It depends on which part of the account you mean. A witness who says they observed an event is presenting a firsthand account of their own observation. A journalist who interviewed that witness has firsthand knowledge that the interview took place, but the event description is attributed to the witness unless the journalist also observed it or independently documents it.

Keep those layers visible. “The journalist spoke with the witness” is not the same claim as “the journalist witnessed the event.” Also distinguish observation from interpretation: someone may directly see a vehicle leave, for example, but an explanation of who was inside or why it left may come from another person or an inference. Record the basis for each important detail rather than assigning one label to an entire article.

Use one row per link in the trail. Here is a hypothetical example; the names and claim are placeholders, not a real event:

1. Post A says: “The library closed early because of a water leak.” Its basis is unclear; it links to Article B.
2. Article B says it got the closure explanation from a post by Account C. Article B is reporting Account C’s statement, not necessarily describing its own observation.
3. Account C says: “I work at the library and saw water in the lobby.” That is a claimed firsthand observation of water in the lobby. It does not, by itself, establish the reason for closing or whether the whole building was affected.

The useful result is a knowledge-and-attribution map: claim; who made it; what they say they observed or consulted; whom they quote or cite; and what remains unconfirmed. If you can open the referenced account or document, check whether it actually supports the wording passed along. If the trail ends at an unnamed or inaccessible source, mark it “not traceable from what I can see,” rather than filling the gap with a guess.

Treat proximity as information about access, not a credibility score. A direct witness may have had a limited view, may be uncertain about timing, or may be describing only one part of an event. A reporter or later account may add useful context, but you still need to see what is attributed and what is independently supported. The map shows the route of a claim; it cannot prove the claim true.

A reusable note can be as simple as: “Claim: [specific statement]. Source: [person, post, or publication]. Connection: [observed / interviewed someone / cited a record / unclear]. Earlier source: [name or not identified]. What this supports: [narrow description]. What remains unknown: [specific gap].” Keep the wording narrow, and label your own inference as an inference. If the original material is unavailable, say so and avoid presenting the repeated version as independently confirmed.