Which of the following does NOT represent one of the common uses of correlation coefficients?

A. Trying to predict someone's level of income based on their level of education.

B. Gathering evidence that a measurement device delivers consistent results over time.

C. Looking for the likely causes for changes in one variable by systematically changing a second variable.

D. Determining if a new measurement tool really measures what it's intended to measure by comparing its results against another accepted measurement tool.