I am no expert on American political campaigns, and do not know the literature on political adverstisements. I have, however, done a fair amount of qualitative research aimed at measuring the meaning of things in a reliable, replicable way. So I’m curious to know who is using such a method to keep track of “negativity” in campaign ads?

Someone must be. Because the candidates both argued tonight not just that their opponent’s ads are perceived by others to be negative (a poll-based description of people’s impressions rather than the ads themselves) or that their opponent’s ads actually are negative (a subjective claim they can just support anecdotally) but that they know exactly how negative their opponent’s ads are.

Obama claimed that John McCain’s ads are “100% negative.” (Does he mean each ad is 100% negative, or that 100% of the ads are at least 1% negative?) Who has coded all of McCain’s ads to determine their negativity according to some reliable rubric?

McCain made even more sweeping claims: “My opponent has run the most negative campaign in history and I can prove it.” This “proof” would require not only demonstrating absolute negativity in ads but coding all comparable ads throughout American history to demonstrate relative negativity.

These are empirical (and empirically falsifiable) statements about the content of ads themselves. But neither cites their source. Who is keeping track of this, and how rigorous, I’m wondering, are the methods used? How does one measure “negativity” in ads such that coders of different political persuasions, working independently, would code the same ad the same way a reasonable amount of the time? What’s the actual definition of a negative ad, and what does the codebook look like? Which candidate came closer to being right on this question?