Methodology
How we evaluate evidence
Supplement marketing routinely cites "clinically studied ingredients" in a way that blurs together very different kinds of evidence. Here's exactly how we sort it out.
Our evidence hierarchy
| Level | Evidence type | How we treat it |
|---|---|---|
| 1 | Human randomized controlled trial | Highest relevance, especially if replicated and the population matches general adults |
| 2 | Systematic review / meta-analysis | Strong when it synthesizes multiple consistent human trials |
| 3 | Small or single controlled human study | Noted as preliminary; we flag sample size, funding source, and replication status |
| 4 | Observational human study | Association only — we do not present it as causal |
| 5 | Animal research | Useful for plausibility and mechanism, not treated as human evidence |
| 6 | In-vitro / lab research | Mechanistic only — we say so explicitly |
| 7 | Traditional/historical use | Context, never presented as clinical proof |
| 8 | Insufficient evidence | We say so plainly rather than filling the gap with a vague claim |
The distinction we never blur
Research on an individual ingredient is not the same as research on Neuro Serge, the finished product. A study on green tea catechins at a specific dose tells you something about green tea catechins at that dose — it does not tell you what a capsule containing an unconfirmed amount of green tea extract, combined with five other ingredients, will do. We flag this distinction on every ingredient page and in the evidence library.
How we source claims
- We prioritize PubMed, PMC, NIH institutes (NCCIH, NIA), and peer-reviewed journals.
- We fetch and read the actual source before citing it — a headline or abstract summary from a search engine is not enough.
- We note study limitations (sample size, funding source, population, duration) rather than only reporting positive results.
- Where we could not verify a claim against a real source, we either remove it or say plainly that we couldn't confirm it.
What we won't do
- We won't cite a study whose topic doesn't match the claim being made (for example, using a paper about sleep or metabolic syndrome to support a brain-health claim).
- We won't present animal or lab research as if it were human evidence.
- We won't imply FDA approval or clinical proof that doesn't exist.
- We won't manufacture statistics, studies, or citations.
See this standard applied across every ingredient in our evidence library.