Just because two things rise together, does one cause the other?
Ice cream sales rise - drownings rise! Does ice cream cause drowning? NO! Summer causes BOTH! This is the #1 statistical error: confusing CORRELATION (things happening together) with CAUSATION (one causing the other). Let's dive deeper!
CORRELATION = two things happening together or following patterns. CAUSATION = one CAUSES the other. Correlation is easy to find - causation requires proof! Just because A and B move together doesn't mean A→B, could be B→A, C→both, or coincidence!
CONFOUNDING VARIABLE = hidden third factor causing both! Ice cream/drowning example: SUMMER is confounding variable (causes both). Shoe size correlates with reading ability - does big feet = smart? NO! AGE is confounding (older = bigger feet AND better reading)!
Maybe B causes A, not A causes B! "Sick people take more medicine" - does medicine cause sickness? NO! REVERSE: sickness causes medicine-taking! "Firefighters are present at fires" - do firefighters CAUSE fires? Always consider direction!
To prove A causes B need: (1) A happens before B, (2) Correlation exists, (3) NO other explanation (ruled out confounders), (4) Mechanism (HOW does A cause B?). That's why controlled experiments matter - isolate the one variable!
Correlation ≠ Causation! Things can move together without one causing the other!
The four possibilities when A and B correlate:
1. A causes B: Smoking → lung cancer ✓
2. B causes A: (reverse causation) Wealth → education? Or education → wealth?
3. C causes both: (confounding) Summer → ice cream AND drowning
4. Coincidence: Nicolas Cage movies correlate with swimming pool drownings (meaningless!)
Famous spurious correlations:
• Divorce rate in Maine correlates with margarine consumption
• Number of people who drowned falling into pool correlates with Nicolas Cage films
• Shoe size correlates with salary (age is confounder!)
Requirements for causation:
✓ Temporal precedence (cause before effect)
✓ Correlation exists
✓ No plausible alternative explanations
✓ Dose-response (more A = more B)
✓ Plausible mechanism (HOW?)
✓ Consistency across studies
Critical question: When seeing correlation, always ask: "What else could explain this pattern?"