How we grade credibility
Every story carries an honest, plain-English read on how much to trust it — a verbal grade, a few evidence badges, and one line of reasoning. We never use a fake star rating or a made-up number. Here's what it all means.
The four grades
- Solid — strong evidence: a well-designed study, a good sample size, and results that clearly support the claim. Reserved for papers we read in full.
- Promising — good evidence with real limits: the finding looks real, but it needs more or bigger studies to be sure.
- Early — a first hint: small, preliminary, or in animals only. Interesting, but a long way from settled.
- Thin — treat with caution: the evidence is weak or the claim runs ahead of what was shown.
The evidence badges
- Preprint — not yet peer-reviewed: posted publicly before other scientists have vetted it.
- N = …: how many people, animals, or samples were studied. Bigger is usually more reliable.
- Double-blind RCT: a randomised controlled trial — the gold standard for testing whether something actually works.
- Mouse study (or another animal): results in animals often don't carry over to people.
- Abstract only: we could only read the paper's summary, not its full methods — so we grade it more cautiously.
"On the fence"
Some papers look striking but leave us with a real doubt we can't resolve. Rather than publish them as if they were settled, or bury them, we list them at the bottom of the edition with a one-line note about what gave us pause — so you can take a look and decide for yourself.