Correlation ≠ Causation
When two things move together, it’s tempting to think one causes the other. But correlation only measures how strongly two variables are related, not whether one makes the other happen.
For example, ice cream sales and shark attack incidents rise and fall together each year. Buying ice cream doesn’t cause shark attacks; hot weather increases both swimming and dessert cravings. A hidden variable, the confounder, can explain both.
Learning to spot these relationships helps avoid bad decisions: like mistaking a marketing bump caused by a holiday for proof that a new ad campaign worked. Correlation gives clues, not conclusions.
Explore an interactive visualization of ice cream vs shark attacks
When the Numbers Line Up by Accident: Not Every Pattern Has Meaning
Tyler Vigen’s Spurious Correlations project famously shows absurd examples like “divorce rate moving almost perfectly with per-capita margarine consumption.” It’s a fun reminder that data can be misleading, and hilarious.
Worth Reading
If you want to dig deeper into why our brains love false cause-and-effect stories, Daniel Kahneman’s Thinking, Fast and Slow is essential. It explores how intuition often jumps from correlation to causation, and how to slow down and think statistically.