Two Ways to Study People
Health research mostly comes in two forms:
- Observational studies watch what people already do — who takes a supplement, who doesn't — and compare their outcomes. They can follow huge groups for years.
- Randomized controlled trials (RCTs) assign people to an intervention or comparison at random, then measure what happens (see [What Is an RCT?](/learn/what-is-an-rct)).
The difference is decisive: observational studies can show that two things travel together (a correlation); only an RCT can show that one *causes* the other.
Confounding: The Core Problem
A confounder is a hidden third factor that explains an apparent link. Classic example: people who take a daily vitamin tend to exercise more, eat better, and see doctors more often. If they end up healthier, was it the vitamin — or the lifestyle that comes with being the kind of person who takes vitamins? This 'healthy-user effect' has fooled many headlines.
Reverse Causation
Sometimes the outcome drives the exposure, not the other way around. If people with early illness stop taking a supplement, the supplement can look protective simply because healthier people kept taking it.
When Observational Studies Shine
They aren't second-class — they're essential when an RCT would be impossible: studying rare outcomes, long-term diet, or anything unethical to assign. They're also how many safety signals are first spotted.
Why Supplement Headlines Overreach
Many 'supplement linked to better health' stories come from observational data. The honest reading is *associated with*, not *causes*. Time and again, when randomized trials test a promising observational signal, the benefit shrinks or disappears — which is exactly why the stronger design matters. The U.S. National Library of Medicine and NCCIH both publish guides on reading these claims carefully [1][2].