Polling in the GOP Primaries

Nate Silver has an excellent look back at how pre-primary polling is doing this year. The short answer is that polling a week out is way off, and only in the last three days are the polls really accurate. This sounds right to me:

If I had to use a word to describe the behavior of Republican voters so far in this race, it would be this one: indifferent. Their preferences between the various candidates have been very weak, shifting after seemingly every debate, primary and news cycle. Many Republican voters like two or more candidates equally well — or at least see two or more of them as being less bad than the others.

I want to emphasize the strange position that nomination politics puts voters in. Remember, three years ago most Republicans were inclined to like Rick Santorum, and Mitt Romney, and Newt Gingrich, and Herman Cain, and really the whole field. Not that most voters knew who Santorum and Pawlenty were, but if they showed up on Fox News, generally most Republicans would find them perfectly agreeable politicians.

I'm going to be a bit lazy here and quote from what I said when Newt was surging in November:

Moreover, most rank-and-file Republican voters just don’t care very much about the subtle differences between Romney and Cain and Perry and Gingrich and the rest. They pretty much like ’em all; after all, they’re all basically conservatives, and they’re all Republicans, aren’t they? The one that they’ll pick if they happen to get polled is therefore most likely going to be whichever one they’ve most recently heard something positive about, which in most cases probably boils down to whoever has been in heaviest rotation on Fox News recently. Two months ago that was Prince Herman. Now, it’s Newt. Current poll numbers, in other words, aren’t a good measure of firm decisions about who folks are going to support; they’re just placeholder answers for a question that the overwhelming bulk of Republicans haven’t really thought about much yet.
This doesn’t represent exceptionally irrational behavior on the part of GOP primary voters, either. After all, unless they live in Iowa or New Hampshire, voters won’t ever be choosing from this unwieldy ten candidate (or so) field, and the odds are good that for most Republicans, in most states, they’ll never have to make any choice since the nomination will be wrapped up before it gets to them. So why should they waste their time trying to figure out which one is the “real” conservative? Why go through the painful business of choosing sides? Picking a horse entails (as Democrats might recall from four years ago) finding things to dislike about a politician who you actually have nothing at all against, used to like, and will like again in the future. Not to mention potentially taking sides against your friends and neighbors, instead of agreeing with them about politics as you normally do (since most of us are surrounded much of the time by co-partisans). Much better to avoid the risk of cognitive dissonance and ignore the whole thing or treat it as entertainment until decision time comes. And with any luck, you can skip it and go straight to the part where you get to go back to disliking Barack Obama.

There so much volatility this time around because there's nothing out there anchoring people's views; there's no heavyweight candidate to organize everyone's opinions, for or against, the way that many Democrats had strong feelings about Hillary Clinton in 2007. For most Republican voters, all the candidates this time (save I guess for Ron Paul) start out as Just Another Republican, and stay that way until the intense last-minute blizzard of ads and appearances show up in their state.

As usual, it's important to remember that even for primary voters, who are relatively high-information compared with everyone else, the campaign just isn't all that visible. I know I haven't seen a single TV ad (granted, I don't watch very many TV ads, but still), and I've seen a handful of Ron Paul bumper stickers, and that's about it, here in Texas were the primary isn't soon.

So put it all together, and it makes lots of sense that it would produce the pattern that Silver finds. It's not a polling problem; it's a public opinion problem of trying to measure something that does not yet exist.