Our method
How we grade evidence
Every substance page carries a badge that says what kind of evidence stands behind it — a human trial, a rat study, or a forum post are three very different things, and we refuse to let them blur together. The badge grades the study design; a separate verdict line says what that evidence actually found. This is our own editorial scale, built for one job: making the gap between "proven in people" and "promising in rodents" impossible to miss.
01 Principle
The one idea that makes it honest: two questions, not one
Most health writing smuggles two different questions into a single word like "strong." We split them, always:
- How good is the evidence? — that's the tier badge below.
- What did that evidence find? — that's a separate plain-English verdict: helped, no effect, mixed, harmful, or unknown.
Keeping them apart matters more than it sounds. A well-run human trial that findsno effect is strong evidence — against a use. A rat study with a dramatic result is still animal-only, no matter how exciting the graph. We say both things every time: how good the evidence is, and what it showed.
02 The scale
The tiers, strongest to weakest
We grade a specific use, not a whole compound. The same molecule can be a human-trial-backed weight-loss drug and, for longevity, nothing but a hypothesis — so it gets two different grades on the same page. Here's the scale:
- Human RCTA randomized controlled trial in people. The strongest kind of evidence we grade.
- Human (controlled)A controlled human study, but not randomized — good, with more room for confounding.
- Human observationalPeople were watched, not assigned — cohorts, case series, single reports. Shows association, not cause.
- Animal-onlyTested in rats, mice, or other animals, with no human data. A lead, not a result.
- In-vitroCells or tissue in a dish. Tells us a possible mechanism, nothing about a living body.
- Mechanistic-hypothesisA reasoned story from biology — a shared receptor or pathway — with no direct test of this use.
- AnecdotalSelf-reports, forum threads, testimonials. Personal experience, never proof.
03 Key flag
The most important flag: "no human data"
Cutting across the whole scale is a separate flag —None-in-humans — that fires whenever a use has never been studied in a single human trial of any kind. For most research peptides, that's the honest headline, and it belongs louder than anything else on the page. A compound can be "Anecdotal + None-in-humans" at once: plenty of people have tried it, nobody has studied it. Reddit is human experience, not human study data, and we don't let one stand in for the other.
04 Scope
What the badge is — and isn't
- It's a description, not an endorsement. The badge rates the evidence, not our opinion of the compound.
- It's ours. We were inspired by the principle behind systems like GRADE and Cochrane — tell the reader how good the evidence is — but we use our own labels and thresholds. No outside body reviews or endorses our grades, and we never imply one does.
- It grades the headline use. The badge up top reflects the use people actually search for, so a strong grade earned for some unrelated endpoint can't be borrowed to make an unproven use look proven.
- It never travels alone. A bare tier with no verdict is only half the story, so we don't ship one.
05 Method
How we decide a tier
For each use, we walk down the ladder and stop at the first kind of evidence that exists: a human randomized trial? A controlled-but-not-randomized human study? Human observational data? Animal studies? Cells in a dish? Just a mechanism? Only forum reports? Whatever rung we stop on is the tier — and if we stop below the human rungs, the None-in-humans flag comes with it. A review inherits the tier of the studies it pools: a review of rat studies is still animal-only.
This scale sits inside our broader editorial policy, and it's part of why we can keep the site free of scare-banners — the honesty lives in the grade itself. Educational information, not medical advice.