Correlation vs Causation
The critical distinction between two things occurring together and one actually causing the other.
Also known as: Correlation is not causation, Spurious correlation, Causal inference
Category: Concepts
Tags: statistics, critical-thinking, research, logic, analysis
Explanation
Correlation vs causation is the fundamental distinction between two variables being associated (they move together) and one actually causing the other. Correlation means: when A increases, B tends to increase (or decrease). Causation means: A directly produces B. The problem: correlation is easy to observe, causation is hard to prove. Common traps include: confusing correlation for causation (ice cream sales and drowning both rise in summer - but ice cream doesn't cause drowning; heat causes both), reverse causation (A correlates with B, but B causes A, not vice versa), and confounding variables (hidden third factor causes both). Establishing causation requires: temporal precedence (cause precedes effect), mechanism (plausible explanation), elimination of alternatives (controlling for confounds), and ideally experimental manipulation. Questions to ask: Is there a plausible mechanism? Could causation be reversed? What third factors might explain both? Would intervention on A change B? For knowledge workers, this distinction helps: evaluate research claims critically, avoid acting on spurious patterns, and recognize that 'data shows correlation' doesn't mean 'data proves causation.'
Related Concepts
← Back to all concepts