Scientific Claim Forensics — Q&A
The method, question by question.
Concrete answers about claims, evidence, and the exact points where reasoning tends to outrun what was actually shown.
01
What is Scientific Claim Forensics?
A method for asking whether the evidence actually warrants a public claim. It takes the claim apart into its exact propositions, traces each to its source, and checks whether the step from source to conclusion holds — or quietly overreaches.
02
What is a claim?
A public assertion that something is true — typically that an intervention works, a result generalises, or a product does what it says. Claim Forensics works on the claim as stated, not a charitable paraphrase of it.
03
What is an atomic proposition?
The smallest checkable unit of a claim. One sentence often bundles several: that an effect exists, that it is caused by X, that it applies to everyone. Splitting them apart is what makes a claim testable rather than merely persuasive.
04
What does “supported” mean?
That the cited evidence establishes the specific proposition, at the strength claimed, for the population claimed. Support is not all-or-nothing: evidence can support a narrow version of a claim while leaving a broader version unestablished.
05
Is a citation enough?
No. A citation shows a source exists; it does not show the source establishes the claim. A real, correctly quoted reference can still be the wrong evidence for the specific proposition it is attached to.
06
What is an overclaim?
A conclusion stated more strongly, broadly or certainly than its evidence supports. The source may be genuine and the quotation exact — the gap is between what was shown and what is asserted.
07
What is mechanism-vs-evidence?
A plausible biological or physical mechanism explains why an effect could happen; it is not evidence that the effect does happen, at a meaningful size, in practice. Mechanism motivates a study — it does not replace one.
08
What is correlation-vs-causation?
Two things moving together need not be cause and effect. Without a design that isolates the cause, an association supports “related,” not “A causes B.”
09
What is product transfer?
Assuming an effect shown for an isolated ingredient or dose carries over to a finished product. Formulation, dose, bioavailability and context can each break the transfer.
10
What is endpoint mismatch?
When the outcome that was measured is not the outcome being claimed — a biomarker or surrogate stands in for a clinical benefit that was never directly shown.
11
What is population mismatch?
When a result from one group — a subgroup, an age band, a specific condition — is presented as if it applied to everyone. The finding may be real and still not generalise.
12
What does “not established” mean?
That the available evidence does not warrant the claim as stated — not that the claim is false. It is a statement about the evidence, and it leaves the question genuinely open.
13
What is the difference between “no evidence found” and “no evidence exists”?
“No evidence found” describes a search: nothing turned up where we looked. “No evidence exists” is a claim about the world. The first never proves the second.
14
Can a real quotation still support the wrong conclusion?
Yes — this is the central case. An exact, correctly attributed quotation can be used to support a proposition it does not actually establish. Genuine provenance and sound inference are separate checks.
15
Does INSPEKTORIUM determine absolute truth?
No. It assesses whether a specific claim is warranted by specific evidence under stated boundaries. It does not pronounce on ultimate truth, and every verdict is open to revision.
16
What role does human review play?
Central. Deterministic checks handle provenance and structure; the harder judgment of whether evidence warrants an inference is made with human review, and the method is designed to make that judgment inspectable rather than to remove it.
17
Which domains can Claim Forensics eventually apply to?
The reasoning patterns are general — any field where public claims lean on cited evidence. Today’s work is grounded in scientific and health-adjacent claims because the failure modes there are well understood and consequential.