Business Applications of Multiple Regression

44 BUSINESS APPLICATIONS OF MULTIPLE REGRESSION

Simply stated, if we wish to show that A causes B, simply showing

that A and B are correlated is not enough. However, if A and B are not

correlated, that does show that A does not cause B— that is, the lack of

correlation between foot size and spelling ability in high school is, by

itself, enough to conclusively demonstrate that having larger feet does not

cause a student to spell better.

Three things are required in order to show that A causes B:

1. A and B are correlated.

2. If A causes B, then A must come before B. This is called a clear tem-

porary sequence.

3. There must be no ...