NP: Mike Keneally + Metropole Orkest, The Universe Will Provide

Like nearly everyone else trying to get the overall gist of last week's election, I've been trying to figure out the whole "moral values" deal. That bastion of out-of-step liberal elitism, according to some of that selfsame "everyone else," the New York Times, has an interesting story on the exit polls that teased out that particular factoid.

Like the cellphone polling issue earlier, this resonates a bit for me as a researcher. The GOP response that "moral values" was a catchall for several key issues does a lot more to make the opposition's point, in my view. If it is, in fact, representing several different issues, it may be a qualitatively different answer choice than some of the others presented. There's a cumulative effect. If your number one issue is Iraq, with, let's say, 35% of your "mind share," and your next two are gay marriage at 25% and Janet Jackson's boob at 15%, the latter two combine to be bigger than the former, even though Iraq is the most important as a single issue.

Of course, you might be able to say the same about the economy, as an amalgam of jobs, taxes and the like, but I think those are less disparate. I could try to make a statistical argument here with a bit on how factor analysis works, but it would be really, really boring.

Still, it's never good form to have such vastly different answer choices in your poll questions, because people are going to respond differently to a category of issues than to a single issue. Then again, trying to frame choices as being between entirely disparate options is part and parcel of politics -- choose my local domestic policy or have my opponent let the terrorists win -- so I guess I shouldn't be surprised by this. I just don't have to like it.