Theorem
Independent variables have correlation
Statistics • Correlation and Regression
Formula
If and are independent (and neither is constant), then , so . The converse is false: does not make and independent.
How to use it
- 1Why it works: the correlation coefficient measures how far and move together in a straight-line way. Independent variables have no tendency to move together, so the measure is .
- 2Read 'independent' as an instruction: the correlation is , with no working.
Worked example
The height of a student and the number shown on a die rolled by the student are independent. What is their correlation coefficient?
Step-by-step
- 1Independent, so
Why: Neither variable tells us anything about the other.
- 2
Answer
Watch out
means no linear relation only: on values symmetric about gives , though depends on completely.
Also known as
uncorrelated independent variables, correlation of independent variables, covariance of independent variables is zero