Clicked Gallery

Correlation vs. Causation: What's the Difference?

Highlighted from a real textbook passage. Explained by Clicked.

Used in a sentence

University Course Reader · STEM

The authors caution that the link is a correlation and that the data cannot establish cause.

The reader highlighted one term — in an article or a paper PDF. Clicked explained the statistics term “correlation” in plain language:

Explained in three depths

Same facts, different vibe — Slang mode 😎

The Clicked way

●○○

Overview

A correlation means two things move together; causation means one of them makes the other happen. Finding a correlation is easy and proves nothing on its own, because two things can rise together while a third thing drives both. Almost every overstated headline about a study lives in this gap.
●○○

Overview

Correlation means two things move together. Causation means one of them is doing it to the other. Spotting the first is easy and proves nothing, because a third thing can be quietly driving both, and that gap is where most bad headlines are born. 😎

A quick take — often all you need.

●●○

Detail

Ice cream sales and drownings rise and fall together all year, and nobody thinks dessert is drowning anyone. Summer is driving both. That hidden third factor is called a confounder, and it is the most common reason a genuine correlation does not reflect cause and effect. The second trap is reverse causation. Anxious people may use their phones more, but that does not tell you whether phone use increases anxiety or whether anxious people simply reach for their phones more often. Coincidence is the third: data from 2000 to 2009 shows a correlation of about 94.7% between US cheese consumption and deaths from bedsheet entanglement. Getting to causation takes an experiment that changes one thing on purpose while holding the rest steady. Where that is impossible, it takes years of separate evidence pointing the same way, which is how the link between smoking and cancer was eventually established.
●●○

Detail

The more firefighters show up, the worse the damage, so obviously firefighters wreck buildings. Or, and hear me out, big fires draw more trucks and also do more damage. That lurking third thing is called a confounder and it is the number one reason a real pattern is not cause and effect. Trap two is reverse causation: people who sleep badly are more stressed, sure, but is the bad sleep causing the stress or the stress wrecking the sleep? Trap three is pure coincidence, and these are real: margarine eaten per person tracked the divorce rate in Maine almost perfectly for a decade, and nobody thinks the spread was breaking up marriages. To upgrade to causation you have to intervene, change one thing on purpose and hold the rest still, or spend years stacking up evidence that all points the same way. Which is why the careful paper says associated with, and the press release about that same paper says causes. 😎

Want more? One click digs deeper.

●●●

Analogy

Every time the neighbour washes his car, it rains. He has noticed, you have noticed, and it really does keep happening. But the pattern alone cannot tell you which explanation is right. Maybe washing somehow brings the rain, maybe he washes on clear days that were always going to cloud over, or maybe you both remember the funny coincidences and forget the dry weeks.
Deepest level reached.
●●●

Analogy

Kids who own more books do better at school, so buy every kid a shelf of books, problem solved. Except the shelf is not doing the work: homes with lots of books usually come with money, time, quiet rooms and parents who read. The books are a symptom of the thing that helps, and shipping books alone into a home that has none of the rest changes far less than the graph promised.
Deepest level reached.

Unfamiliar concept? A real-world example makes it click — fresh analogies on tap.

AI explanations may contain errors · Not professional advice

Formal definition — The same term, explained the usual way

Correlation denotes statistical association between two variables, whereas causation denotes that variation in one produces variation in the other. Association may arise from confounding, reverse causation, selection effects, or chance, and therefore does not establish causality; causal inference requires experimental manipulation or, absent that, converging observational evidence assessed against criteria such as temporality, dose response, and mechanistic plausibility.

Want Clicked to explain terms like “correlation” directly in your browser — including on PDFs?

Add to Chrome — Free

50 free Explanations · No credit card required