Comment by rwilson4

4 years ago

If the relationship is linear, I'm guessing a test based on Pearson has greater power.

Seems likely. The presented coefficient only looks at the ordering of the X-values, and how it relates to the ordering of the Y-values. All other information is thrown away. That's how it can be so general, but it should come at the expense of power.

  • Ah that makes sense intuitively. I was confused how non linear correlation could be detected without knowing anything about the function itself.

power and interpretability.

Assuming a linear relationship, if you know the correlation coefficient, you can predict unobserved values of y based on a known x with good accuracy.

y = ax + b + error

where strong correlation means error is small.