Methodology: How We Evaluate the Evidence
Every claim on this site passes through one funnel: what kind of study made it, how strong and consistent that evidence is, and what it can — and cannot — actually tell us. This page is the funnel, in full view.
the funnel, in full view
The Evidence Tiers
| Tier | Label | What counts |
|---|---|---|
| T1 | Strong | Multiple high-quality randomized trials, strong systematic reviews and meta-analyses, established clinical guidelines |
| T2 | Moderate | Consistent prospective cohorts, limited but credible trials, replicated observational findings |
| T3 | Preliminary | Small trials, early human studies, single observational associations |
| T4 | Experimental | Animal, cellular, and mechanistic research — informative about biology, silent about human outcomes |
| T5 | Speculative | Anecdote, influencer claims, marketing without credible research |
What the Star Ratings Mean
- ★★★★★ Strong — the claim survives systematic review and consistent trials.
- ★★★★☆ Good — solid evidence with real remaining questions.
- ★★★☆☆ Moderate — credible associations and some trials; causation not established.
- ★★☆☆☆ Preliminary — early human data or strong animal biology awaiting human confirmation.
- ★☆☆☆☆ Very limited — mostly mechanistic or speculative; the honest answer is "we don't know yet."
Rules Every Topic Follows
- 🔗 Correlation ≠ causation — observational associations are labeled as associations, never as proof.
- 🧬 Biomarkers ≠ outcomes — improving a biomarker is reported as exactly that; the leap to "lives longer" requires outcome data.
- 🐁 Animal evidence is labeled — mechanisms from mice and cells are presented as mechanisms, not human findings.
- 📊 Absolute over relative — where possible, we report absolute risk and effect sizes, not just percentages.
- 📅 Everything is dated — every topic carries "Evidence last reviewed." Conclusions can change as new research lands.
The Pillar Weighting Model
The homepage ranks ten pillars with weights summing to 100% — sleep 20, exercise 15, nutrition 15, metabolic health 10, relationships 10, stress management 10, and the remaining four at 5 each. These weights are editorial judgment: our read of evidence volume, effect size, and the strength of causal support. They are not measurements, and we say so on the page. A pillar at 5% is not unimportant — hormetic stress and purpose simply carry less settled evidence than sleep and exercise. The weighting is a prioritization tool, not a precision claim.
The Review Process
- 🕵️ Research — primary sources first: trials, systematic reviews, guidelines, then cohorts; secondary coverage only for context.
- ⚖️ Cross-reference — multiple independent sources; conflicting evidence is reported, not buried.
- ✍️ Write — with the voice rules in the Editorial Standards page: cautious language, uncertainty on the table, risks beside benefits.
- ✅ Review — every topic passes an automated QA gate (structure, links, chart geometry, banned hype lexicon) before it's shown to readers.
🧭 The guiding question
The site optimizes for accuracy over engagement. When the evidence is weak, the correct output is "we don't know yet" — and you'll find that sentence on these pages more often than on most health sites. That is a feature, not a bug.
The Bottom Line
- Five evidence tiers — from systematic reviews down to speculation, labeled on every claim.
- Stars reflect evidence quality — not enthusiasm, not popularity.
- Weights are honest editorial judgment — a map, not a formula.
- Everything is dated — and corrected when the evidence moves.