Correlations indicate a relationship, but no cause/effect relationship between the two or more items being studied.

A good example might be the realtionship between the number of doctors and the number of blacksmiths in America between 1900 and 2000.

They show a nearly perfect inverse correlation--the rate of rise of one number is mirrored by te fall of the other number.

If we were to derive a cause/effect from such data we could conclude that balcksmiths were becoming doctors. Being reasonable though, we know that a third factor has actually generated the change in both numbers--something totally unrelated directly to either doctors or blacksmiths.

Correlational relationships are not in any way empirical evidence of anything other than the ability to utilize certain mathematical formulae.
President Reagan summed it up nicely, "Some people go through their entire lives wondering if they made a difference, Marines don't have that problem!!!"