How We Source Evidence

What counts as evidence here

We rank sources by study design. Randomised controlled trials and systematic reviews of them sit at the top. Observational studies, mechanistic work, and animal studies are labelled as such and are never used on their own to support a recommendation.

Where we look

Primary literature through PubMed and Cochrane, position statements from national bodies, and drug and safety databases. We read the abstract, then the methods section. A press release is not a source.

What we record for each study

Sample size, duration, dose used, the form of the ingredient, who was studied, and what was actually measured. A trial on sixty men aged twenty to thirty does not generalise to women over fifty, and we say so rather than quietly widening the claim.

How we handle dose

When we state a dose we state the dose used in the trial and the form it was given in. Magnesium oxide and magnesium glycinate are not interchangeable at the same milligram figure, and we do not write as if they were.

When we say we do not know

Often. If the trials are small, short, or contradictory, the honest answer is that the evidence is not there yet, and we write that instead of splitting the difference.

Corrections

If you find a figure we got wrong, email editorial@nutrimosaic.com. Verified errors are corrected and logged on the corrections page.