Sources & methodology.
How we decide what counts as evidence — and how we cite it.
Source hierarchy
- Peer-reviewed research (PubMed-indexed journals)
- Published clinical trial results (journal publications and trial registry records)
- Regulatory and pharmacopeial documents (FDA, USP, EMA)
- Named expert commentary in journals or conference proceedings
We do not cite forums, vendor marketing, anecdotes, or social media as evidence. When community discussion points to a topic worth covering, we write about the underlying research — not the thread.
How we evaluate a study
- Design — randomized and controlled beats observational; replication beats novelty.
- Scale and duration — we note sample sizes and follow-up lengths, because "in mice for six weeks" and "in 2,500 humans for two years" are different claims.
- Population — results in one population don't automatically transfer to another.
- Conflicts — we note funding sources when they're material to interpretation.
How we cite
Claims link to their sources inline — usually PubMed records — and longer guides end with a references section. Where we can't verify a claim against a primary source, we don't make it.
Handling conflicting findings
Where studies disagree, we describe the disagreement rather than picking the convenient side, and we weight by design quality and replication. "The research is mixed" is an acceptable conclusion.
Limits of our coverage
We cover published research and testing standards. We don't test products ourselves, and we don't evaluate vendors beyond the documented, verifiable signals in our sourcing guides.