Correlation Is Not Causation

THE MENTAL MODEL: Just because two events happen at the same time, or two numbers move in the same direction, does not mean one is causing the other. An observed association is a starting point for investigation, not a conclusion. Mistaking correlation for causation is a logical fallacy known as 'cum hoc ergo propter hoc' (with this, therefore because of this).

HOW IT WORKS: When you see a correlation between two variables, say A and B, there are several possibilities. First, A could cause B. Second, B could cause A. Third, and most commonly overlooked, a third factor C could be causing both A and B to happen. This third factor is called a confounding variable. For example, an increase in user engagement (A) might correlate with a new feature release (B), but both could have been caused by a holiday season (C). Finally, the correlation could be pure coincidence.

WHEN TO USE IT: Apply this critical thinking lens anytime you analyze data. This includes interpreting dashboards, reviewing A/B test results, debugging performance issues, or assessing the impact of a new initiative. Always ask: 'What else could be causing this? Is there another factor at play?' The goal is to move from observing a correlation to designing an experiment that can prove or disprove a causal link.

WHEN NOT TO USE IT: You should never ignore this principle, but don't let it lead to inaction. A strong correlation is often a valuable signal that warrants further investigation. While it doesn't prove causation, it helps you form a hypothesis. The mistake isn't noticing the correlation; it's stopping there and not digging deeper to find the true causal relationship through controlled experiments or other statistical methods.

ONE CANONICAL EXAMPLE: A famous example is the strong positive correlation between ice cream sales and crime rates. It would be absurd to claim that buying ice cream causes crime, or that committing crimes makes people crave ice cream. The confounding variable is hot weather. When it's hot, more people are outside, leading to more social interactions and, unfortunately, more opportunities for crime. Simultaneously, more people buy ice cream to cool down.