Causality vs Correlation

Also known as · Correlation vs Causation

A regression slope shows association; a causal reading needs the other factors held fixed. A single-regressor model on observational data leaves confounders in the error.

When to use

Assignment 1 asks "is this causal?" twice and the answer is no both times: mothers who smoke or people who study differ in other ways that also affect the outcome, an Omitted Variable Bias problem.

Worked in Assignment 1.

Taught in Lecture 1.

Appears in