Card 07

📊 🔗 ❓

Ice cream sales and drowning deaths both rise in summer. Does ice cream cause drowning?

💭 How to Think About This

Obviously not! But this silly example reveals a serious trap. Two things moving together (CORRELATION) doesn't mean one causes the other (CAUSATION). A hidden third factor—hot weather—causes BOTH. This mistake leads to wrong conclusions in medicine, policy, business, and everyday thinking.

When two things rise and fall together, what's your first instinct?

🔒 Write your thinking above to unlock hints

Assuming correlation = causation is the MOST common reasoning error. When A and B move together, we instinctively think one causes the other. But there are actually FOUR possibilities: A causes B, B causes A, something else causes both, or pure coincidence. Which is it here?

A CONFOUNDING VARIABLE is a hidden third factor affecting both things. Hot weather → more swimming → more drowning. Hot weather → more ice cream. The weather causes BOTH. Ice cream and drowning are correlated but neither causes the other! Can you spot confounders in other examples?

This trap leads to bad decisions: "Breakfast eaters are healthier!" (Maybe healthy people choose breakfast) • "Hospital patients die more!" (Sick people go to hospitals) • "Music lessons boost grades!" (Rich families can afford both). Seeing correlation, assuming causation = dangerous.

To PROVE A causes B, you need: (1) Randomized experiments (assign A randomly, see if B changes), (2) Mechanism (plausible explanation for how A leads to B), (3) Time order (A happens before B), (4) Dose response (more A → more B). Correlation alone proves nothing!

No—hot weather causes BOTH ice cream sales and drowning. They correlate but don't cause each other!

Your instinct to see causation is the exact trap this problem exposes. It's the most common reasoning error.

Four possibilities when A and B correlate:

• A causes B

• B causes A (reverse causation)

• C causes both A and B (confounding)

• Pure coincidence

Famous confounding traps:

• Fire trucks and fire damage (bigger fires cause both)

• Shoe size and reading ability (age causes both)

• Storks and babies (rural areas have both)

The protection: When you see correlation, don't ask "which causes which?" First ask: "Could something ELSE cause both?"

You spotted the CONFOUNDING VARIABLE trap! When A and B move together, a hidden third factor (C) often causes both. In this case: Hot weather → more swimming → more drowning. Hot weather → more ice cream. Weather is the confounder causing the misleading correlation.

When A and B correlate, there are exactly four possibilities: (1) A causes B, (2) B causes A, (3) C causes both A and B (what you spotted!), or (4) Pure coincidence. Without careful investigation, you can't know which! Your instinct to suspect a third factor is scientifically mature.

Your confounder-hunting skill applies broadly: "Music lessons → better grades" (family values cause both) • "Hospitals → more deaths" (illness causes both) • "Expensive schools → success" (wealth causes both). The hidden third factor is usually socioeconomic, demographic, or environmental.

Researchers use RANDOMIZATION to eliminate confounders: randomly assign people to treatment/control groups. This ensures confounders are equally distributed. Observational data (just watching correlations) can't prove causation—only experiments can, because they control for hidden third factors.

Exactly right! Hot weather causes BOTH—that's the confounding variable you suspected!

You showed sophisticated probabilistic thinking by looking for the hidden third factor.

The four possibilities when things correlate:

• A causes B (rare to confirm without experiments)

• B causes A (reverse causation)

• C causes both (confounding—what you spotted!)

• Coincidence (spurious correlation)

Your confounder-hunting habit is invaluable:

• "Linked to" → What's the third factor?

• "Associated with" → What causes both?

• "Correlated with" → What's the confounder?

The principle: Never assume correlation is causation. Always hunt for the hidden variable causing both.

Pure coincidence DOES happen! With enough data, random correlations appear everywhere. "Nicolas Cage films correlate with swimming pool drownings"—obviously meaningless! But the ice cream/drowning case has a REAL connection through weather. It's not coincidence—there's a hidden cause.

This isn't coincidence—it's a CONFOUNDING VARIABLE. Hot weather → people go swimming → some drown. Hot weather → people buy ice cream. The correlation is REAL, just not causal. Something else (weather) causes both effects. This is more common than pure coincidence!

COINCIDENCE: No real connection, random pattern (Cage films/drownings). CONFOUNDING: Real connection through hidden third variable (ice cream/drowning via weather). The difference matters! Confounded correlations persist; coincidences don't. Always check for that hidden cause.

To distinguish coincidence from confounding: Is there a PLAUSIBLE mechanism connecting them? (Weather connects ice cream and swimming—yes!) Does the correlation PERSIST across different times/places? (Random coincidences usually don't.) Is there a clear third factor? Always hunt for the hidden cause.

Not quite coincidence—there's a real connection through HOT WEATHER!

You were right to be skeptical of causation, but this isn't pure randomness. It's confounding.

The distinction matters:

• Coincidence: No real connection (random)

• Confounding: Real connection through third variable

• Ice cream/drowning is confounded, not coincidental

The mechanism:

• Hot weather → swimming → drowning

• Hot weather → ice cream sales

• Weather causes both—real connection, no causation

The habit: When you see correlation, ask: "Is there a third factor causing both?" That's more common than pure coincidence.

🤔 Which thinking lens(es) did you use?

Select all the lenses you used:

Learn more about the 7 Thinking Lenses →

👨‍👩‍👧 For Parents & Teachers

🌱 A Small Everyday Story

"Students who take music lessons get better grades!"
said the headline.
Parents rushed to enroll their kids.
But wait—maybe families who value education
both sign up for music AND help with homework.
The music might not be the cause at all.

See more guidance →

🧠 Thinking habits this builds:

Automatically asking "but does A actually cause B?" when seeing correlations
Hunting for hidden third factors (confounders)
Understanding why randomized experiments are the gold standard
Being skeptical of headlines that imply causation from correlation

🌿 Behaviors you may notice (and reinforce):

"But couldn't something else cause both?" questions
Noticing when news confuses correlation with causation
Asking about the direction of causation
Looking for mechanism explanations, not just associations

How to reinforce: When you encounter a "X linked to Y" claim, play the confounder game together: "What third thing might cause both X and Y?"

🔄 When ideas are still forming:

Some learners may become overly skeptical and think we can never know if anything causes anything. Help them understand that causation CAN be established—through experiments, mechanisms, and careful analysis—it just requires more evidence than mere correlation.

Helpful response: "Correlation is a clue, not proof. What additional evidence would convince you there's real causation here?"

🔬 If you want to go deeper:

Explore the website "Spurious Correlations" for absurd examples
Discuss how randomized controlled trials establish causation in medicine
Look up Simpson's Paradox for an advanced confounding puzzle

Key concepts (for adults): Correlation, causation, confounding variable, reverse causation, spurious correlation, randomized controlled trial, selection bias, Simpson's Paradox.

Ice cream sales and drowning deaths both rise in summer. Does ice cream cause drowning?

🔄 Compare All Perspectives

🔗 "One causes the other"

🔍 "Something else causes both"

🎲 "Could be coincidence"

🤔 Which thinking lens(es) did you use?