The statistics tag appears to be a popular meta-tag, which is to be be avoided.
Should it be deleted and banned?
I'm probably one of the "guilty" users, but I find the tag useful to note questions where the key evidence is or should be about the interpretation of statistical data. I don't think this makes it a meta-tag. Perhaps I just need to see a better explanation of why using it is a problem, or perhaps a better, clearer explanation of the concept of meta-tag.
I think we would be worse off without it. It clearly distinguishes some questions from others, though it applies to a wide range of questions. Perhaps it is a unique problem on this site because, to some degree, it classifies the sort of evidence required to address a question not just the content of the question.