Ice cream sales and drownings rise together. Nobody drowns because of ice cream. Today’s skill is the one the exam rewards most in this unit: saying exactly what the data supports — and refusing to say more.
A real correlation question on real data. Before reading the answer, decide what you’d need to know. Enrolled in our Runestone course? Open it from there so your progress counts.
Big Data, at its most important edge: the limits of a claim. Everything so far produced numbers. Now: what are those numbers allowed to mean?
Two variables are correlated when they tend to change together. Causation means one actually produces a change in the other. Correlation is what data shows easily; causation is much harder to show.
The usual culprit is a confounding variable: a third factor that drives both. Hot weather sends people to buy ice cream and sends people swimming. Ice cream and drownings move together because summer moves both.
| When you see a correlation, ask | Because |
|---|---|
| Is there a third factor driving both? | confounding variable |
| Could it run the other way? | reverse causation |
| Did people choose which group they’re in? | self-selection: the groups differed before anything happened |
| Could it be chance? | small samples produce patterns that don’t last |
Twelve months of ice cream sales vs drownings (made-up numbers, realistic shape). Each dot is a month. Then turn on the hidden variable.
S = supported by the data as stated · X = not supported (it claims cause from correlation) · ? = need more information.
Five questions. Pick the most cautious answer that’s still true.
“X linked to Y” becomes “X causes Y” somewhere between the study and the share button. Real decisions — school policies, medical advice, laws — get made on that jump. Being the person who asks “what else could explain this?” is not being difficult. It’s being right.