How to Evaluate a Claim: A Practical Checklist

Investigation5 min readPublished October 8, 2026

Most of us meet dozens of surprising claims every day: a post says a food causes cancer, a video says a government is hiding something, a headline says a new study overturns everything. We cannot check them all, but we can build a habit of asking a few good questions. This checklist works for almost any claim, and it is the same set of questions that investigators and scientists use.

1. What exactly is being claimed?

Vague claims are hard to test. Before judging anything, restate it in one sentence: who did what, when, and with what result. If you cannot do that, the claim may be too fuzzy to check, and that is already a warning sign. "Scientists say coffee is dangerous" is vague. "A 2021 study of 10,000 adults found that people who drank four cups a day had a 5 percent higher rate of a given condition" is testable.

2. Where did it come from?

Trace the claim back to its original source. A viral post usually cites a news article, which cites a press release, which summarises a study. Each step can lose detail. Aim for the primary source: the paper, the court record, the data set, the original footage. If nobody in the chain links to it, be careful.

3. Who is saying it, and why?

Everyone has a perspective. Ask whether the source has relevant expertise, whether they have something to gain, and whether they have a record of being right or wrong. A financial interest does not make a claim false, but it is a reason to look for independent confirmation.

4. What is the evidence?

Distinguish between types of evidence:

Anecdotes: stories about one person or event. They are vivid but cannot show how common something is.
Observational studies: they find patterns in data, but cannot easily prove cause.
Controlled experiments: the strongest tool, especially when randomised and repeated.
Physical evidence: documents, photos and objects, which need to be verified themselves.

A claim backed by several independent kinds of evidence is stronger than one backed by many repetitions of the same kind.

5. Is there a simpler explanation?

Occam's razor suggests that among competing explanations, we should prefer the one that makes the fewest unsupported assumptions. A strange light in the sky can be an alien craft, or an aircraft, a planet or a lens reflection. The more ordinary explanations should be ruled out first, not because they are more exciting, but because they are more often right.

6. Could it be tested and shown wrong?

The philosopher Karl Popper argued that a good scientific claim must be falsifiable: there must be some possible observation that would show it to be false. A claim that explains every possible outcome, so that nothing could ever count against it, is not really telling us anything. If a believer says that any missing evidence is proof of a cover-up, that is a sign the idea cannot be tested.

7. How big is the claim?

The saying that extraordinary claims require extraordinary evidence was made popular by the astronomer Carl Sagan, who built on earlier versions from the sociologist Marcello Truzzi. The point is not that unusual things never happen, but that the weight of evidence should match the size of the claim. A report of a rare but ordinary event needs a little proof; a claim that overturns well-established physics needs a great deal.

8. Correlation is not causation

Two things happening together does not mean one caused the other. Ice cream sales and drownings both rise in summer, and the cause of both is warm weather. When a claim says "X causes Y", ask whether there could be a third factor, or whether the cause might run the other way.

9. Check the numbers

Look for the denominator. "Doubled" could mean from one in a million to two in a million. Ask about sample size, how people were chosen, and whether a result was one of many tested. When a study tests twenty things, one will often look significant by chance.

10. Watch your own mind

We are all prone to confirmation bias, noticing evidence that fits what we already believe and discounting what does not. A useful test is to ask, "What would change my mind?" If the answer is nothing, then the belief is not really based on evidence.

A quick worked example

Suppose you see a post that says "a miracle supplement cures fatigue, doctors hate it." Applying the checklist:

The claim is vague. Which kind of fatigue, and how much improvement?
The source is probably the seller, who has a financial interest.
The evidence is likely testimonials, which are anecdotes.
The simpler explanation is that people feel better with time, or that expectations matter.
The size of the claim, a cure, is large, and the evidence is small.

None of this proves the supplement does nothing. It tells you that the claim has not earned your trust yet.

Practice

The best way to learn is to practise on real cases. Our Case Room takes famous mysteries and asks you to weigh the evidence yourself, and our articles on the moon landing, Roswell and the Bermuda Triangle show how these questions apply to well-known claims.

Published October 8, 2026. Written with AI assistance and prepared for publication under our editorial policy. Spotted a mistake? Tell us and we will check it.