Correlation vs Causation Quiz
Ice cream 'causes' drowning, firefighters 'cause' damage, red wine 'extends' life... Twelve scenarios test whether you can spot the confounder, the backwards arrow, or pure chance behind a correlation — and recognize the one causal claim that actually holds up.
0 / 12 answered
Q1
A beach town notices that in months when ice cream sales rise, drowning deaths rise too. A local paper runs the headline 'Ice cream is killing our swimmers.' What is the most likely explanation?
Q2
A running club finds that members who wear knee braces report far more knee injuries than members who don't. Someone proposes banning braces to protect knees. What's the flaw?
Q3
Across an elementary school, students with bigger feet score higher on reading tests. Should the school buy everyone bigger shoes?
Q4
A rooster crows every morning shortly before sunrise, without fail. The farm's youngest child concludes the crowing makes the sun come up. Which reasoning error is this?
Q5
A study app claims it raises exam scores. Which design would give the strongest evidence that the app actually causes higher scores?
Q6
A school survey finds that students with higher self-confidence earn better grades. The principal launches a confidence-building program to raise grades. What's the strongest objection?
Q7
A fan wears his 'lucky jersey' for five matches and his team wins four. He now refuses to wash it, insisting the jersey drives the wins. What's the best assessment?
Q8
City data show that the more firefighters sent to a fire, the greater the property damage. A columnist argues the city should send fewer firefighters. What's wrong with this?
Q9
A basketball fan swears that when a player gets 'hot hands' and hits several shots in a row, the next shot is more likely to drop. What did the most famous research on this actually find?
Q10
Hospital A has a lower overall cure rate than Hospital B. Yet for mild cases A beats B, and for severe cases A also beats B. How can this be?
Q11
Several randomized trials assign insomniacs either to a 'no screens for an hour before bed' group or to their usual habits, and the no-screen groups consistently fall asleep faster. A friend shrugs: 'Correlation isn't causation.' Is the shrug justified?
Q12
A headline reads: 'Study finds people who drink a glass of red wine daily live longer.' Before toasting to your health, which question should you ask first?
Answer all 12 questions to see your result 👆
Cheat sheet: the vocabulary of causal traps
- Confounding variable
- A hidden third factor that drives both variables, creating a correlation with no direct causal link (summer heat behind ice cream sales and drownings).
- Reverse causation
- The arrow points the other way: B causes A, not A causes B (sore knees lead people to wear braces, not the reverse).
- Randomized controlled trial (RCT)
- The gold standard for causation: random assignment makes the groups comparable in every respect, so outcome differences can be attributed to the treatment itself.
- Post hoc fallacy
- 'After this, therefore because of this' — treating mere temporal order as causation (the rooster and the sunrise).
- Simpson's paradox
- A trend that holds inside every subgroup can reverse when the groups are merged, because the groups differ in size and makeup (the two hospitals).
- Natural experiment
- When life randomizes for you: a policy change, a lottery, or a border creates comparable groups without a lab.
- Multiple comparisons
- Test enough variable pairs and some will correlate strongly by pure chance; a 'discovery' dredged up this way needs fresh data to confirm.
Correlation vs causation: the difference, the traps, and how to think straight
| Correlation | Causation | |
|---|---|---|
| Definition | Two variables move together: when one is high, the other tends to be high (or low) | A change in one variable actually produces a change in the other |
| Visible in raw data? | Yes — a scatter plot or a correlation coefficient reveals it directly | No — the same data are compatible with confounding, reverse causation, or chance |
| Who predicts whom | As long as it's stable, either variable predicts the other, whatever drives what | Answers the intervention question: change the cause, and the effect follows |
| How it's established | Simple observation of the data is enough | Gold standard is a randomized controlled trial; observational data need confounders controlled and reverse causation ruled out |
| Classic fail | Ice cream sales track drowning deaths (confounder: summer) | Concluding the rooster's crow raises the sun because it always comes first |
Correlation means two things move together: ice cream sales and drowning deaths both peak in July; shoe size and reading scores climb side by side across an elementary school. Causation is a far stronger claim — that changing one variable will actually change the other. The difference between correlation and causation is the difference between noticing a pattern and understanding a mechanism, and nearly every statistical howler in the news comes from blurring the two.
Whenever you meet a correlation, three rival explanations must be ruled out before you believe 'A causes B'. First, a confounding variable: some third factor drives both (summer drives ice cream and swimming alike). Second, reverse causation: B causes A instead (sore knees cause brace-wearing, not braces causing sore knees). Third, plain coincidence: with small samples, or enough variable pairs tested, striking correlations surface by chance alone.
So how is causation actually established? The gold standard is the randomized controlled trial: random assignment makes the groups alike in every respect except the treatment, so any difference in outcomes must come from the treatment itself. When experiments are impossible — you can't randomize people into smoking — researchers combine statistical control of known confounders, natural experiments, temporal order, dose-response patterns and plausible mechanisms, building a case file rather than delivering a single knockout proof.
Everyday life is a minefield of causal overreach. Health headlines turn 'wine drinkers live longer' into drinking advice while ignoring that wine drinking tracks wealth and healthcare access. Wellness posts credit a supplement for recoveries that would have happened anyway. And investors canonize whichever fund manager beat the market five years running, forgetting that among thousands of managers, somebody's streak is guaranteed by chance.
This is also prime standardized-test material. Causal flaws — mistaking correlation for causation, overlooking a confounder, flipping the arrow — rank among the most frequently tested argument flaws in LSAT logical reasoning and are a staple of GRE argument analysis. If you're prepping, the instincts you trained here transfer directly: try our LSAT logical-flaw quiz and GRE argument quiz next.
The way to improve is not to memorize the slogan 'correlation does not imply causation' but to run the checklist until it fires automatically: Who claims A causes B? Could a third factor drive both? Could the arrow point backward? How big is the sample? Was anything randomized? Twelve questions from now, that checklist will feel less like homework and more like a reflex.
Frequently asked questions
Does correlation ever imply causation?
Correlation alone never proves causation, but it isn't worthless — it's usually the first clue in a causal chain. A correlation becomes genuine causal evidence when it's paired with temporal order, a plausible mechanism, ruled-out confounders and, ideally, randomized experiments. 'Correlation isn't causation' means 'don't stop here', not 'nothing can ever be known'.
What are the three main explanations for a correlation besides causation?
A confounding third variable that drives both sides (summer heat behind ice cream sales and drownings), reverse causation where the arrow points the other way (injuries cause brace-wearing), and coincidence — small samples and multiple comparisons throw up striking but meaningless correlations by chance.
What are some real examples of correlation without causation?
Textbook classics include ice cream sales and drowning deaths (both driven by summer), children's shoe size and reading ability (both driven by age), the number of firefighters at a fire and the damage done (both driven by fire severity), and the rooster's crow 'causing' the sunrise (mere temporal order).
How do scientists prove causation?
The gold standard is the randomized controlled trial: random assignment equalizes the groups, so outcome differences can be attributed to the treatment. Where experiments are impossible, researchers combine statistical control of confounders, natural experiments, dose-response patterns, temporal order and mechanisms to build the causal case step by step.
Is correlation vs causation tested on the LSAT or GRE?
Heavily. Mistaking correlation for causation is one of the most common flaws in LSAT logical reasoning questions, and GRE argument essays almost always contain a causal leap waiting to be dismantled. Practicing on examples like the ones in this quiz is direct test prep.