Data distribution in an XML document

Is there any information available on the distribution of data within an XML document? i.e. how much data is encoded in attributes and how much data is encoded in element content?

I understand this would vary greatly across domains - I would presume in the document centric world [i.e. publishing, document repositories, email archives, discussion boards, etc] much of the content would be part of the elements through mixed content, etc rather than attributes - but I was wondering if any work had been done in analyzing the distribution of data with an xml document.

thanks,

prakash

---

Do you Yahoo!?
All your favorites on one personal page � Try My Yahoo!