Oh -- I forgot to ask an important question: what happens when you compress the PDF? If you can compress the PDF so it more closely equals the word document in size, the practical difference between them sort of evaporates, at least as to storage size.
To me the advantage of a PDF over a Word document is that the former is now open-source. And it is universally regarded as the "correct" format for technical writing.
> ... I might actually share this idea of your of compressing political speeches to identify the uniqueness in what is said in the speech ;-)
The intriguing part is that it's an objective measure, in a field that's called "political science" but isn't remotely scientific. I'd like to see a thorough study comparing political speeches in this easily conducted, almost automatic way.
Another, perhaps more difficult, study would analyze advertising copy for use of words that have no objective meaning, words that appeal only to emotion. My favorite advertising word is "zesty", a frequently-used word with no information content whatever.