Correlation vs. Causation: The Mistake Behind Headlines

"People who drink coffee live longer." "Poor sleep linked to lower income." Headlines like these quietly invite you to hear cause where the study only found company. It's the single most common error in health reporting — and once you can see it, you can't unsee it. Here's how to catch it every time.

Short Answer: Correlation means two things tend to occur together; causation means one actually produces the other. Most health studies — and nearly all the headlines about them — find correlation, then imply causation. The gap matters because acting on a false cause wastes effort or backfires. A few questions reliably tell the two apart.

Why This Matters

If you learn one thing from this pillar, make it this. The confusion of correlation with causation underlies a huge share of misleading health claims, and spotting it is the highest-leverage science-reading skill there is. [1]

Correlation means two variables move together. Causation means one actually brings about the other. Correlation is necessary for causation but nowhere near sufficient — and most studies can only measure the former. [1]

Science Explanation

Confounding is the usual culprit: a hidden third factor drives both measured things. People who take a supplement may also exercise, sleep, and eat differently — so a link between the supplement and health may belong to the lifestyle, not the pill. [2]

What Research Shows

Observational studies — which simply watch what people already do — can reveal strong, real correlations but generally can't establish cause, because the groups differ in countless unmeasured ways. This is exactly why they sit mid-ladder in the evidence hierarchy, below randomized trials that can isolate cause. [1]

To test a causal-sounding claim, ask: Could it run the other way (reverse causation)? Could a third factor explain both (confounding)? Was this a controlled trial or just observation? Has it been replicated? If a claim implies cause from a single observational study, treat the causal language as the writer's, not the data's. [1]

The stakes are practical: act on a false cause and you spend effort, money, or hope on something that can't deliver. Calibrating to what the study actually showed — association, not proof — protects both your decisions and your trust. It's the discipline behind every careful claim on this site. [1] Updated 2021 reporting standards for systematic reviews emphasise transparency in literature search methods, inclusion criteria, and risk-of-bias assessment as the key determinants of a review's reliability — providing a structured framework for evaluating any published evidence synthesis. [3]