The Two Independent Sources Rule
MinutesDo not treat a consequential claim as established until two sources that obtained the information separately both support it, and check that they really are separate, because ten outlets repeating one origin are still one source.
How to do it
- 1For each source, ask "how do they know?" Look for the attribution: their own reporting, a document, a named witness, or "according to" someone else.
- 2Follow each chain upstream. If both reports lead back to the same press release, the same wire story, the same anonymous post, or the same single study, count them as one.
- 3Prefer independence of access over independence of brand: two people who each saw the document, two datasets collected separately, two laboratories. Different logos on the same material do not count.
- 4Prefer sources that would pay a price for being wrong (a named reporter, an institution with a corrections policy, a court filing under oath) to sources that lose nothing.
- 5Scale to the stakes. Casual interest needs less; moving money, making a health decision, accusing a person, or sharing with a large audience needs the full rule.
- 6Until the second source arrives, hold the claim as "reported by one source" and say so when you mention it.
What to say
- “Everyone is reporting it, but everyone is citing the same post. Who has confirmed it on their own?”
- “I will treat that as unconfirmed until someone with separate access says the same thing.”
When to use it
- •Breaking news, leaks, and rumours about named people or companies.
- •A statistic or "study shows" claim that many sites repeat.
- •Investment tips, product-safety scares, and health claims spreading through forwards.
- •Any decision you could not easily reverse if the claim turned out to be false.
Counters
Evidence and how strong it is
Independent corroboration is a core norm of professional journalism, described by Kovach & Rosenstiel (2001) as part of the "discipline of verification" that distinguishes journalism from other communication, and it is built into newsroom verification procedures (Silverman 2014). Its logic is statistical: two errors are unlikely to coincide only when they are independent, which is why circular reporting (outlets citing one another back to a single unverified origin) defeats it. The psychological need for the rule is well evidenced: repetition alone increases perceived truth (Pennycook, Cannon & Rand 2018), so hearing a claim from many directions feels like confirmation even when it is one claim echoed. Evidence strength: professional norm and logical argument; experimental support exists for the illusory-truth problem it addresses, not for the rule as an intervention.
- Two sources can share a bias, a funder, or a mistaken upstream dataset and still look independent. Ask what would have to be true for both to be wrong at once.
- The rule sets a floor, not a proof. Two confirmations of a claim are weaker than one primary document you can read yourself.
- Some true things have only one possible source (a lone whistleblower, a single witness). Then the questions become how that source knows, what they risk by speaking, and whether the checkable details hold up.
- Kovach, B., & Rosenstiel, T. (2001). The Elements of Journalism: What Newspeople Should Know and the Public Should Expect. CrownThe account of verification, including independent corroboration and transparency about sourcing, as the defining discipline of journalism.
- Silverman, C. (Ed.) (2014). Verification Handbook: A Definitive Guide to Verifying Digital Content for Emergency Coverage. European Journalism CentrePractical newsroom procedures for corroborating claims and tracing them to an original source.
- Pennycook, G., Cannon, T. D., & Rand, D. G. (2018). Prior Exposure Increases Perceived Accuracy of Fake News. Journal of Experimental Psychology: General, 147(12), 1865-1880Experimental evidence that mere repetition raises belief, the reason echoed claims feel corroborated when they are not.