Supplements · lesson 4 of 11
Evidence — how to weigh a claim
Mechanism vs RCT vs meta-analysis, and what makes a study trustworthy.
7 min read · reviewed October 2026
Every supplement comes with claims. Learning to grade them — to tell 'plausible idea' from 'proven in humans' — is the most powerful skill in this whole track, because it lets you ignore most of the noise.
The key distinction: a mechanism (a reason something might work) is not the same as evidence that it actually does in real people. Plenty of ideas that look beautiful on paper, or work in a petri dish or in mice, fall apart in human trials.
The number of participants in a single landmark supplement RCT (the VITAL trial of vitamin D and omega-3) — the scale needed to settle a question one small study or a mechanism never can.
Source: VITAL randomized controlled trial
The evidence ladder — from weakest to strongest
A plausible biological reason it could work. Necessary but weak on its own — most mechanisms never pan out in practice. Treat 'supports X pathway' as a hypothesis, not a result.
Effects in cells or mice. Useful for discovery, but the doses are often absurd relative to a human, and biology differs. 'Cured cancer in a dish' is the classic overhyped headline. Promising, not proof.
Tracks groups of people over time. Can reveal associations but can't prove cause — people who take a supplement often differ in many other ways (the 'healthy-user' effect).
The gold standard: people randomly assigned to supplement or placebo, ideally double-blind. Randomization cancels out confounders, so a difference can be attributed to the supplement itself.
Pool many RCTs to see the overall picture and whether results replicate. The strongest single source — but only as good as the trials inside it.
What makes an individual study strong
Trust a study more when it's big, randomized, blinded, uses a real dose for long enough, isn't funded by someone selling the product, and has been repeated by other teams.
How it works →
Randomization balances unknown confounders across groups so the only systematic difference is the treatment. Blinding (ideally double-blind, where neither participant nor researcher knows who got what) prevents the placebo effect and unconscious bias from creeping in. Adequate sample size gives the statistical power to detect a real effect and reduces false positives. A dose and duration matching real-world use ensures the result is actually relevant. Independent funding and pre-registration reduce the chance results were cherry-picked.
What the studies show →
A useful real example of evidence done well is the VITAL trial — a large, randomized, placebo-controlled study of vitamin D and omega-3 in tens of thousands of adults. It's a model for how a definitive supplement question gets answered, and its nuanced results (benefits for some endpoints, not others) show why a single mechanism or small study is never enough. The broader lesson: weigh the body of evidence and its quality, not the loudest single claim. Where high-quality replicated RCTs exist, lean on them; where only mechanism or mouse data exist, stay skeptical.
A supplement ad cites a study showing its ingredient 'killed cancer cells in the lab.' How much should that move you?
- A lot — that's strong proof it works
- A little — it's an early signal that may or may not hold up in humans
- Not at all — lab studies are worthless
- It depends only on how big the effect was
Show the answer →
B.A little — it's an early signal that may or may not hold up in humans
Cell and animal results are how research starts, not how it ends. Many compounds that destroy cells in a dish do nothing useful (or are harmful) at realistic human doses. It's a hint to investigate, never a reason to buy. Demand human RCTs for human claims.
Each ingredient in the encyclopedia carries an evidence summary, so you can judge the strength of a claim before you spend anything.