Google tests a new feature that shows the approximate number of words next to certain search results. The word count is usually displayed for articles, but only for a small number of results. You'll probably think twice before clicking on a search result that has 50,000 words, unless you're looking for an in-depth report.
The idea that Google spends resource recording things like word count - or even more, reading level - intrigues me. I appreciate that this might just be research for now, rather than being used in the ranking algorithm. But it doesn't stretch the imagination too far to foresee matching those factors with Google's "intention engine" data.
As a rough example, a transactional intention might depress pages with very long copy - but an informational or scholarly intention might lean toward longer and higher reading level content.