I recently read this really good article about the meaning of p-value. I believe it highlights one of our many misunderstandings about statistics.
In the same vein, there is this paper entitled "Why Most Published Research Findings Are False" and published in PLOS Medicine. The statistical argument is not easy to follow but it highlights the fact that when we analyze our own data, we are sometimes testing tens or thousands of different correlation without ever correcting for multiple comparisons. What is then the meaning of a correlation with p-value of .049? Additional evidence for the claimed relationships are then most than welcome.