The Claim TestSubscribe

The Claim → The Evidence → The Verdict

Popular claims.
Actual evidence.

Every day, we're surrounded by claims about health and science.

The Claim Test asks a simple question: what does the evidence actually support?

We examine the research, challenge the conclusions, and give each claim a clear, calibrated verdict.

The Weekly Test

One claim. The evidence. The verdict. Once a week.

Free. No spam. Unsubscribe anytime.

What a Claim Test looks like.

The Claim

“Magnesium helps you sleep.”

Verdict

Limited evidence

Some randomized trials suggest magnesium may improve certain sleep outcomes, but the studies remain relatively small and the observed benefits are generally modest. The evidence is not yet strong enough to say magnesium reliably improves sleep.

What supports it

Randomized trials have reported improvements in some sleep outcomes.

What weakens it

Study sizes are limited, results vary by outcome, and objective sleep measures have not consistently improved.

What we can say

Magnesium may help some people, but a promising signal is not the same as a proven sleep aid.

Sources shown on every published Claim Test.

A simple test. A rigorous process.

  1. 01

    The Claim

    We start with a specific, testable claim people are hearing online, in the news, or in everyday life.

  2. 02

    The Evidence

    We examine the strongest available research, including study quality, effect size, limitations, conflicting findings, and uncertainty.

  3. 03

    The Verdict

    We translate the evidence into a conclusion that says no more — and no less — than the science supports.

AI helps us search.
A scientist makes the call.

Modern AI can search, organize, compare, and challenge scientific information at remarkable speed.

We use those capabilities to make evidence review more efficient — not to outsource scientific judgment.

Before publication, every Claim Test verdict is reviewed by a human scientist.

  1. Search
  2. Synthesize
  3. Challenge
  4. Human review
  5. Publish

AI can propose a verdict.
It cannot approve one.

Scientific review

Julien Dimastromatteo, PhD

Biomedical scientist and scientific editor with experience evaluating research across molecular imaging, oncology, translational science, and drug development.

His role at The Claim Test is simple: challenge the evidence, calibrate the conclusion, and make sure every published verdict says no more than the science supports.

Not everything is true or false.

Science rarely fits into a binary answer. The Claim Test uses five verdicts to show how strongly the available evidence supports a claim.

Strong
Consistent, high-quality evidence supports the claim.
Moderate
The evidence supports the claim, but meaningful uncertainty remains.
Mixed
Credible evidence points in different directions or depends strongly on context.
Limited
There are promising signals, but the evidence is not strong enough for a confident conclusion.
Unsupported
The better available evidence does not support the claim, or credible evidence contradicts it.

One claim worth understanding.
Once a week.

Get The Weekly Test — a short evidence review that separates what we know from what we think we know.

Free. No spam. Unsubscribe anytime.