Check out a very similar and interesting study on the word choice of 08 presidential candidates

<link>

That study provides another really interesting way to measure speech and tone quantitatively. James Pennebaker, a social psychologist at the University of Texas, has yet a third method.

<link>

Hopefully, some ambitious politics thesis-writer will do a meta-analysis so we can see whose results align the most with our intuition -- or debunk the conventional wisdom. In the mean time, I think we can all agree that the Times method of simply counting word frequency, while it makes for neat-looking graphics, is pretty primitive.