Ed Kilgore has been doing some great blogging today (here and here) about a new Third Way poll of supposedly “swing independent” voters.

I have two things to add. One is methodological. Third Way defines their “swing” independents as those without strong views of either Barack Obama or Mitt Romney. Given that Romney’s approval ratings are in the dumps these days pending an almost certain post-primary rally among strong conservatives, it seems to me that their group is likely to contain quite a few Republican voters who haven’t yet admitted, even to themselves, that they’ll wind up strongly supporting Romney in the fall. I couldn’t tell from the release, but it looks to me, in other words, that their polling suffers from the problem that Jonathan Chait identified last week: If the universe of committed voters supports one candidate, then that candidate doesn’t have to break even, much less win, among undecideds.

The second point is actually a big one about polling in general. What the math of survey research tells us is that we can confidently say, if the mechanics are done correctly, that a very large group of people would respond similarly to the questions asked of a very small group of people. That can be useful! But it says nothing at all about what the response actually means. If it’s fairly simple — say, the day before the election, asking people who they plan to vote for — you can get pretty useful responses. However, the more abstract the questions are, the less they are easily interpreted. So, for example, Third Way’s conclusion that “Swing Independents think we should fix the deficit over reducing income inequality” promises a lot more than it can deliver. After all, most people have no idea what “fix the deficit” really means, other than it’s generally thought to be a Good Thing. Which is, in fact, all that the survey is telling us. Whether any actual votes would switch if Barack Obama talked more about the deficit, let alone actually proposed something that would slash the deficit (which, of course, requires either raising taxes or cutting popular spending programs or both), is another story altogether.