Correlation is the most abused means of statistical analysis in existence. Everyone repeat - most correlations are spurious and/or non-causal. Pay little attention to them unless followed up by sound methodology in causality.
How could you know? IMHO, our ability to intuitively correlate, often with very imperfect knowledge, is what makes us great guessers. I think most human correlations (where there is smoke, there is fire) are likely to be correct. However, we also need to be humble and we need to try to begin to understand our limitations and our tendencies toward hubris and self-deceit.