Separate correlation from causation
Tell 'two things move together' from 'one thing causes the other,' and spot a hidden third factor (confounder).
- Reading a trend between two variables: Correlation is a trend between two variables, so you first need to read whether they rise or fall together.
- Controlled comparison: Causation is shown by holding other factors constant, so you need the idea of comparing groups that differ in only one way.
Prerequisites are inferred: pending teacher review.
Re-learn the skill with worked practice and clear examples.
A correlation means two things move together; causation means one makes the other happen. A confounder is a hidden third factor that can create a correlation without any direct cause.
Across towns, more streetlights are linked with more nighttime crime. A reporter writes 'streetlights cause crime.' What is the most likely flaw?
Reviewed- A.Streetlights truly cause crime
- B.Town size is a confounder: bigger towns have both more streetlights and more crime
- C.The two variables are not correlated
- D.Crime is an outlier
Show the worked solution ▾
Answer: B. Town size is a confounder: bigger towns have both more streetlights and more crime
- Step 1: Check the claim type: The reporter jumped from 'linked' (correlation) to 'cause' (causation) with no test.
- Step 2: Find a hidden third factor: Bigger towns have more of almost everything: more streetlights and more crime: so town size can drive both.
Why it's right: Town size is a confounder tied to both variables, so the correlation does not show that streetlights cause crime.
- A: A correlation alone cannot prove this cause.
- C: They are correlated; that is given.
- D: This is about a confounded correlation, not a single stray point.
Aligned to Biomedical Innovation: correlation vs. causation · reading level ~grade 9
- A news graph showing two rising lines is often a confounded correlation, not proof of cause.
Fill these in as you work through the lesson.
- Correlation (two things tend to move together):
- Causation (one thing actually makes the other happen):
- Confounder (a hidden third factor linked to both):
- Controlled comparison (groups that differ in only the one thing you test):
When two things rise together you have a , but that is not proof of ; before claiming cause, rule out a hidden third factor called a .
- Ice-cream sales and pool drownings both rise in summer. Name a hidden factor that could drive both.
- Why does 'two things move together' not prove one causes the other?
- What kind of study would let you actually test whether a factor causes an effect?
Over one summer, neighborhoods with more parks also have lower asthma rates. Explain why this correlation does not prove parks cause less asthma, and name a possible confounder.