# Evidence Levels

Every effect on this wiki gets two separate ratings. One says how sure we are. The other says what the trials found. A confident "no benefit" is a strong result, not a weak one.

## How sure we are

| Evidence | What it takes |
|----------|---------------|
| Strong   | A meta-analysis, or at least two independent randomized trials, measuring the outcome itself (not a blood marker), with consistent results in a population like the one named |
| Moderate | Randomized trials exist, but one problem lowers confidence: small samples, short duration, some inconsistency, or a narrow population |
| Limited  | One or two small trials, surrogate outcomes only, observational data, or trials that disagree |
| Insufficient | Animal, cell, mechanistic or anecdotal evidence only, or no human trials |

These roughly follow the GRADE levels used in clinical research (high, moderate, low, very low).

Caps:

* If every positive result comes from manufacturer-funded trials, the grade is at most Limited.
* When an intervention can't be blinded (sauna, cold water, light) and the outcome is subjective (mood, stress, how recovered someone feels), the grade is at most Moderate.
* Trials shorter than 8 weeks on a long-term claim lose one level.

## What the trials found

Benefit, small / Benefit, moderate / Benefit, large / No benefit found / Mixed results / Not tested.

"Mixed results" means adequate trials pointing in opposite directions. A single unreplicated trial is Limited evidence, not Mixed.

## Where effects go

An effect is listed under **What the research supports** only if trials found a benefit and the evidence is Moderate or Strong. Everything else goes under **Claimed but unsupported or contested**.

## Why there is no single score per article

A supplement can have strong evidence for one effect and none for five others. One overall grade would average things that don't belong together, and it would read as a verdict on the product. Instead, each article opens with an evidence summary: the best-supported effect and the most-cited weak claim, each with its own rating.

Every rating names the population studied, gives a short reason for the grade, and cites at least one source. Observational associations are never described as causes. Harms are graded and cited the same way as benefits.