Evidence

How we know what we know.

Every claim is sourced. Every recommendation is conditional. Every confidence label is honest. This page is our methodology, published.

The citation rule.

We do not state a thing about biology that is not traceable to a study, a meta-analysis, or a regulatory document. This is not a style preference. It is the operating constraint. If we cannot cite a claim, we do not make it.

When the evidence is absent or inconclusive, we say so plainly: “the research on this is limited,” or “this has not been adequately studied in the relevant population.” We treat “we do not know this yet” as a complete and acceptable answer.

Where a compound or practice is discussed speculatively, the confidence label is Speculative and the language is conditional throughout: “clinical studies suggest,” “animal models indicate,” “discuss with your physician before considering.”

Source tiers.

We cite at four tiers, in order of evidential weight. The tier is disclosed wherever it materially affects interpretation.

TIER 1

Highest weight

Systematic reviews and meta-analyses of randomized controlled trials (Cochrane, JAMA, NEJM, Lancet). We cite these as strong evidence when they exist.

TIER 2

Substantial weight

Individual randomized controlled trials with adequate power, pre-registered, peer-reviewed. We cite these as good evidence, with sample-size and population caveats noted.

TIER 3

Moderate weight

Observational cohort studies, case-control studies, prospective epidemiology. We cite these as suggestive evidence and note that causality cannot be established.

TIER 4

Context only

Animal studies, in-vitro research, mechanistic hypotheses, expert commentary, case reports. We cite these only to illustrate mechanism or open questions — never as evidence of clinical effect in humans.

Confidence labels.

Every article and protocol summary carries one of three confidence labels. These are editorial assessments, not clinical ratings. They change when the evidence changes.

Established

Multiple Tier 1 or Tier 2 sources in general agreement. Safe to discuss without heavy qualification.

Emerging

Tier 2–3 evidence with meaningful but not yet replicated findings. Discussed with conditional language throughout.

Speculative

Tier 3–4 evidence, or mechanism only, or early human signal. Interesting but not yet actionable — discussed as hypothesis.

What we will never cite.

Knowing what not to cite is as important as knowing what to cite. The following categories are excluded from the evidence base used by this site.

—Podcasts, YouTube videos, or social media posts — regardless of the speaker's credentials.
—Supplement brand white papers, in-house research not independently replicated, or proprietary clinical data not published in peer-reviewed journals.
—Case reports as evidence of general efficacy.
—Animal studies as evidence of human clinical effect (cited for mechanism only).
—Single studies contradicted by the weight of the literature — cited, if at all, as minority position.
—Any source that cannot be traced to a primary document.

Errors and updates.

Evidence changes. We expect to be wrong about some things currently presented as established. When new research materially revises a prior position, we update the affected content and note the revision date. We do not quietly correct without acknowledgment.

If you believe a specific claim on this site is unsupported or misrepresents the literature, write to us. We read these messages. The email is on the Contact page.