debateee

False Cause Fallacy: Mistaking Correlation for Causation

The false cause fallacy treats something as the cause of an outcome when it is not. It includes four common errors: confusing correlation with causation, getting the direction of cause and effect backwards, missing a third factor that drives both, and mistaking coincidence for a link. This guide explains each, with examples and a method for testing causal claims.

Updated 2026-11-04

What is the false cause fallacy?

The false cause fallacy, in Latin non causa pro causa, “not the cause as the cause,” attributes an effect to something that did not produce it. It is a broader category than the post hoc fallacy, which concerns time order only. It covers all ways in which a causal claim can go wrong without enough support.

Causal claims are among the most important in public debate, because they are what justify action: if X causes Y, then change X to change Y. They are also hard to establish, which is why this fallacy is so common and why errors can be costly.

The pattern to look for

Four patterns produce most false cause errors.

  1. Correlation as cause: A and B move together, so A causes B.
  2. Reverse causation: A and B are linked, but B causes A.
  3. Common cause: a third factor C causes both A and B.
  4. Coincidence: A and B line up by chance, especially when many comparisons are made.

Examples in everyday talk and debate

The same error appears in different settings. These examples show the pattern and what is missing from each.

Four types of false cause
TypeClaimBetter explanation
Correlation as cause“Cities with more police have more crime, so police cause crime.”Cities with more crime hire more police (reverse causation).
Reverse causation“Happier people have more friends, so friends make you happy.”Both may be true: friendship and happiness feed each other.
Common cause“Ice cream sales and drownings both rise in summer, so ice cream causes drowning.”Hot weather increases both.
Coincidence“Two unrelated yearly figures follow the same curve for a decade.”With enough variables, some will match by chance.
Oversimplified cause“The company failed because of one bad decision.”Many factors usually contribute.
Confounding“People who take vitamins are healthier, so vitamins improve health.”Vitamin takers may exercise more and smoke less.

Why it is persuasive

We look for causes, and when two things appear together a causal story forms quickly. The stories are often plausible, and plausibility is easy to mistake for proof. Media headlines compound the problem: “Study links X to Y” is often read as “X causes Y.”

Also, causes are rarely visible. We observe results, not mechanisms, so we fill in the gap with what seems likely. Our intuitions are good for simple physical causes and poor for social and biological systems, where many factors interact.

Watch it explained

A short video can make the pattern easier to spot the next time you hear it.

Watch: The danger of mixing up causality and correlation (TEDx Talks (Ionica Smeets)). Real examples of patterns that suggest the wrong cause.

Watch: Critical Thinking: Fundamentals: Correlation and Causation (Wireless Philosophy). Why two things moving together need not be linked.

When it is not a fallacy

Correlation is not worthless. It is often the first clue to a real causal link, and a strong, consistent correlation across settings is a reason to investigate. Smoking and lung cancer were first noticed as a correlation, then supported by evidence of mechanism, dose-response, and many studies. The point is not that correlation never matters, but that it is not enough alone.

Also, claims of causation are not fallacious just because the cause is partial. “Smoking causes lung cancer” does not mean every smoker gets it. A cause can raise the probability of an outcome.

How to respond

A structured reply is stronger than “correlation is not causation.”

  1. Say what the claim is and what evidence is given.
  2. Offer a specific alternative: reverse direction, third factor or chance.
  3. Ask what comparison or experiment would separate the explanations.
  4. Ask about the mechanism: how would A produce B?
  5. Suggest how much weight the claim deserves given the evidence: “suggestive, not conclusive.”

Using this in a debate

Policy debates turn on causal claims, such as “this program reduced crime” or “this tax reduced investment.” The best evidence comes from comparisons: a similar region that did not have the program, or data before and after with a control. If you cite a study, say what kind it is: a randomized trial, a natural experiment, or an observational study. If the other side cites a correlation, ask for the alternative explanations they have ruled out.

Be careful with your own wording. “Linked to” and “associated with” are accurate for correlations; “causes” and “leads to” claim more. Matching your language to your evidence makes you more credible, and it makes it harder for a rebuttal to land.

How researchers build a causal case

Researchers use several kinds of evidence together. Randomized experiments assign people by chance to a treatment or a control, which removes most confounders. Natural experiments use events that happen to split similar groups. Dose-response patterns show that more of the cause brings more of the effect. Mechanisms explain the steps between cause and effect. Consistency across places and times shows the link is not a fluke. No single piece is decisive, but together they can be. A speaker who explains which of these supports their claim is giving the audience a way to judge it.

Related fallacies

These are easy to confuse with this one, or often appear alongside it.

  • Post hoc: assuming that what follows is caused by what came before.
  • Single cause fallacy: ignoring that most outcomes have many causes.
  • Confounding: a hidden factor that distorts a link.
  • Texas sharpshooter: finding patterns after the fact and treating them as predictions.

Frequently asked questions

What is the false cause fallacy?

It is the mistake of treating something as the cause of an effect when it is not, such as confusing correlation with causation or reversing cause and effect.

What is the difference between correlation and causation?

Correlation means two things vary together. Causation means one produces the other. Correlation can arise from cause, reverse cause, a common cause, or coincidence.

What is an example of a false cause?

Ice cream sales and drowning deaths both rise in summer, but ice cream does not cause drowning. Hot weather is a common cause.

How do you prove causation?

Use controlled or natural experiments, show a mechanism and a dose-response pattern, rule out alternative explanations, and replicate the result.

Is post hoc the same as false cause?

Post hoc is one kind of false cause, based on time order. False cause is the broader category that includes confounders, reverse causation and coincidence.

Sources and further reading