Which statement best distinguishes correlation from causation?

Which statement best distinguishes correlation from causation?

• A. Correlation proves causation
• B. Causation implies correlation but not vice versa
• C. Correlation implies causation with a mechanism
• D. Correlation indicates association, causation requires mechanism and ruling out confounding

Answer: (D)

The key idea is that correlation measures an association between two things, while causation requires evidence that one thing actually produces a change in the other. Two variables can move together for many reasons, even if one isn’t causing the other. Correlation tells you there is a relationship, but it doesn’t show why that relationship exists. For example, ice cream sales and sunglasses sales often rise together in summer, but that doesn’t mean buying sunglasses makes people eat more ice cream or vice versa—the warm weather drives both. To claim causation, you need a plausible mechanism that explains how one variable would bring about a change in the other, and you must rule out confounding factors that could create the appearance of a causal link (factors that influence both variables). That’s why the statement describing correlation as indicating association and causation as requiring a mechanism and ruling out confounding best captures the distinction. The other statements misstate the relationship: correlation does not prove causation; causation does not automatically imply a simple, universally guaranteed correlation; and correlation doesn’t inherently imply causation even with a proposed mechanism.

The key idea is that correlation measures an association between two things, while causation requires evidence that one thing actually produces a change in the other. Two variables can move together for many reasons, even if one isn’t causing the other.

Correlation tells you there is a relationship, but it doesn’t show why that relationship exists. For example, ice cream sales and sunglasses sales often rise together in summer, but that doesn’t mean buying sunglasses makes people eat more ice cream or vice versa—the warm weather drives both. To claim causation, you need a plausible mechanism that explains how one variable would bring about a change in the other, and you must rule out confounding factors that could create the appearance of a causal link (factors that influence both variables).

That’s why the statement describing correlation as indicating association and causation as requiring a mechanism and ruling out confounding best captures the distinction. The other statements misstate the relationship: correlation does not prove causation; causation does not automatically imply a simple, universally guaranteed correlation; and correlation doesn’t inherently imply causation even with a proposed mechanism.