Believe Your Lyin' Eyes

Making a prediction about this election boils down to one simple question. What do you believe: your lyin' eyes or the lyin' polls?

I'm going with my eyes.

There are lots of factors, of course, but here's as good a place as any to start: Obama is going to venues that McCain was using in 2008 -- smaller, indoor locations. But where is Romney holding his events now? In the vast, outdoor sites that can hold overflow crowds of tens of thousands -- the ones used by Obama in 2008. Not as relevant but still interesting is that my sister, who lives in deep blue Boulder County in Colorado, says that she has seen more Romney than Obama signs, far more GOP signs that she has seen in a decade of living there.

Why does that matter? Because Obama won in significant part by achieving a D+7 turnout (39% Democrats vs. 32% Republicans) -- the best Democratic turnout since Reagan realigned party ID into near-parity in 1980. With enthusiasm clearly stronger on the right than left, there's just no conceivable way that Dem turnout will even approach D+7 -- but that's the figure that pollsters are using to show Obama tied or slightly ahead.

There are two broad polls of party self-identification, one from Gallup and the other from Rasmussen. Gallup is showing a tiny GOP advantage this year of just one point (36 to 35) -- a swing of 11 points from the 10-point Democratic advantage of 39 to 29 right before the 2008 election. When you add in "leaning" independents, the GOP advantage grows to 3 points, 49 to 46. Looking at Rasmussen's monthly party ID polling, the GOP advantage is a staggering 5.8%, 39.1% to 33.3%. It is the biggest lead for Republican self-identification in the nine years of monthly results posted at the Rasmussen site -- by far.

If turnout is simply close to the dead-even 37-37 turnout from 2004 -- when George W Bush had a turnout operation that was better than McCain's but significantly inferior to Romney's -- then Romney gets 52% of the popular vote, which is such a huge margin that it would mean he'd get at least two of the three of OH, WI, and/or PA, perhaps all three.

Forgetting the topline numbers from the various polls, the polls are actually fairly consistent. Look at the subsets: Romney is winning roughly 95% of Republicans, and Obama is getting 90% -- 95% of Democrats. Most of the polls have Romney winning independents by high single digits or more. CNN's 49-49 national poll on Sunday, as has been widely discussed, had Romney winning independents by 22 points -- roughly the same margins the GOP had with independents in 2010. So how could it have been tied at the topline numbers? As many have noted, CNN had a sample that was D+11 (40-29) -- 4 points better than Obama's best-in-a-generation Dem turnout in 2008.

But there's one number that really sticks out: 15%. That appears to be the percentage, according to various polls that have asked, of Obama 2008 voters who are voting Romney. (Here's the most recent, which shows 13%, which means probably close to 15% by the time the last "undecided" 2008 voters, well, decide.) That means one out of every seven Obama voters are fleeing. Polls say roughly 5% of McCain 2008 voters are voting for Obama, but I honestly have not met a single McCain voter who has said to me, "You know, I was wrong about Obama. He's proved to be much better than I thought." I don't believe it will be a statistically significant number of actual McCain 2008 voters who switch to Obama -- and there simply aren't enough new Democratic-leaning voters who sat out during the historic turnout from four years ago (and NOW are suddenly excited enough to come vote) to make up for all those defectors. (And that's not counting a likely spike in religious Christian voters compared to four years ago.)

Extra evidence that there is genuine Obama "buyer's remorse," look at newspaper endorsements. Almost three dozen editorial boards have flipped from Obama in 2008 to Romney today. Just three have gone from McCain 2008 to Obama. Editorial boards are obviously not right-of-center as a group, and if anything (as a generous assessment), they would be considered left-of-center independents voters. That's real movement -- and a lot of it. Romney, for what it's worth, has earned roughly 60% of the swing state newspaper endorsements (24 to 15, according to They have a little influence, but more important is what they indicate: disillusionment with Obama is real, and it's leading a lot of his 2008 voters to swing to Romney.

Assuming the polling subsets looking at Obama 2008 voters are correct -- and there's a higher probability of that than pollsters correctly predicting relative turnout -- then Team Obama has to discover and turn out millions of new voters who are somehow more excited today than when seemingly the entire known world was giddy over Barack Obama. But how many people are there who were under a rock four years ago, but now are ginned up?

So maybe it's not about lyin' polls, given that the subsets of Republican, Democratic, and Independent voters seem to be relatively consistent across the various surveys. It's what the pollsters are doing to "weight" the various subsets.

So perhaps, then, the question really is: Are you believing your lyin' eyes or the lyin' pollsters?