I love this usage.
Lies, damn lies, and…
Proofiness – Charles Seife – NYTimes.com.
Correlation is not causation. However, you would expect causation to require correlation. So, if one sees no correlation between two measurements, it would seem to imply that either the wrong observations are being made or there is no causal relationship. For example, if conservatives claim that GDP will rise if taxes are lowered, then they should be required to show some instances of when GDP rose when taxes were lowered.
The moral is not that statistics is not useful, but that it’s mostly useful to disprove things, and much less useful to prove things. Proof is hard.