Tags are meant to collect questions on a single subject, so a tag that applies to several completely different kinds of questions is a bad tag.
The statistics tag is one of the worst offenders. I've seen people apply it to:
- Questions on "standards and conventions specific to statistics as an academic discipline, and programs leading to a degree in this field." For example, Should I do a PhD or Master in Statistics for a career in Data Science?
- Questions on "the use of statistics in academic research. Statistics is the science of collecting, analyzing and drawing inference from data." For example, Should I cite all R packages I used? This is the use currently specified by the tag wiki excerpt, which reflected the actual use of the tag at the time that I last edited it, but doesn't seem to now.
- Questions asking for statistics about research or academia. For example, Can you get statistical data on research paper submissions and publications?
I believe the best way to disambiguate is to have differently-named tags to address these different use cases. The tag names should clearly indicate which usage the tag is referring to, since people don't always read the tag wiki excerpt (they're not easy to get to while posting a question on mobile site, for example).
(I believe tags are limited to 24 characters on this site.)