03 June 2010

Correlations Have Causes

The classic expression of skepticism about statistical evidence is that "correlation is not causation."

But, people often forget the converse of this observation which is that "correlations have causes."

Data can show that two things are related, which leaves the researcher with the next step of merely show why that is the case. There is a cause out there to find, and additional research need only discern which of the possible causes it could be.

No comments: