Gravar-mail: How to quantify the evidence for the absence of a correlation