Andrew et. al.'s book is chock-full of facts. Indeed, I bet it has more charts than pages. And yet, I have to confess, I don't feel I came away from the book informed. And I worry that others won't as well.

Gelman and his co-bloggers over at The Monkey Cage (an indispensable blog about political science) are often complaining about irresponsible journalists telling stories at odds with the facts. To pick an example at random, here's Adam Nagourney claiming "Republicans voters have displayed a zeal for their candidates that Democrats could only envy." The bloggers quickly run the numbers and show that, no, the numbers have been pretty similar over the years.

And yet...

The image of zealous Republicans is much stronger than the charts I saw on the blog. Indeed, the image is so strong that we're inclined to say the numbers must be wrong -- perhaps they don't account for [insert your favorite objection here]. And so we repeat the mistaken image again.

This is nothing more than what George Lakoff talks about when he says "the truth will not set you free" -- we have models in our head of how the world works and persuading people requires getting them to switch models, not just giving them facts that prove their existing model is wrong. But, in general, this doesn't seem to be something political scientists are very good at.

The overwhelming tendency among the profession is to just run the numbers. Want to see whether campaign contributions influence votes? Well, you just get a database of campaign contributions, get a database of votes, and do the math. Calling up some donors or some
politicians to ask them about it rarely even seems to occur to them.

Interviews have their flaws, of course -- everybody lies, after all. But numbers have their flaws too. We need both. Without running the numbers, we can't weed out the lies and urban myths. But without interviews, we get the facts without any real picture of why the facts are the way they are. And without that picture, it's hard for people to really grasp what's going on, let alone think of ways to change it.

Thomas Frank's book was powerful because it laid out a very simple picture anyone could understand: working-class Kansans were hoodwinked by religious nutjobs into voting against their own interests. Larry Bartels and Andrew Gelman and others have poked holes in this picture to be sure, but what do they have to put in its place? If you want to win a battle of ideas, numbers aren't going to be enough.