Joe Biden and a Possible Coming Poll Correction

Well, that was quite a debate. Check out Taegan Goddard's post-debate reaction: "The vice presidential debate between Joe Biden and Paul Ryan was one of the best debates I can remember. It was a great service to all Americans." I could've done without all the lying (#LyinRyan was trending on supposedly conservative dominated Twitter last night), but close enough for horseshoes and hand grenades. The consensus among the punditry is that Biden won. A CBS flash poll of undecided voters backs that up. But the consensus is also that it probably won't move polling in Obama's direction. "Biden stemmed the bleeding," seems to be the most overused metaphor of last night and this morning.

But can a vice presidential debate change the dynamics? The consensus seems to be no, but the fact is that anything's possible.


[Nate Silver:]

Vice-presidential debates rarely move head-to-head numbers between the presidential candidates – even when there is a much clearer verdict in instant-reaction polls. So one should err on the side of caution in assuming that the debate had much influence either way.

There is a plausible hypothesis, however, that some of Mr. Romney’s recent surge in the polls reflects a growing “enthusiasm gap” between Democrats and Republicans. To the extent that Mr. Biden’s performance re-energized Democratic partisans, he may have left President Obama in a slightly better position than where he started the night.


And polling-wise, we'll probably see the president's numbers improve -- which likely would've happened anyway. But it's possible that they may improve dramatically. And that's where things get interesting.

See, part of the problem with a post-event bounce is that a certain portion of the numbers are completely inaccurate. Is there this big group of Americans who wander back and forth between candidates like a dog being coaxed with two bones? And is that big group so skittish that bad news sends them scampering off to the other candidate? Probably not. Bounces have a lot of noise.

When pollsters called after the first presidential debate, Obama backers were demoralized and Romney backers were energized. Where the Romney group were probably more likely to answer a poll, the Obama group probably didn't want to talk about the presidential race at all. So part of the sudden swing in support was illusory. It's an unrepresentative spike in the data. After a while, this lack of balance evens out and the bounce comes back down again.

Now consider, the consensus seems to be that Biden reenergized Obama voters and helped close the enthusiasm gap. Will this help erase that statistical noise sooner? Logic would seem to suggest so. We'll have to see. But there is some evidence that the race was basically unchanged by the first debate and that "evening out" would look like a dramatic gain for Barack Obama.

Keep in mind, this is all speculation at this point. I'm just saying what might happen. But none of this strikes me as completely unreasonable or even unlikely. And what's important here isn't the fact that it happens, since it's artificial in the first place, but that everyone takes it the wrong way if it happens -- as they did with Mitt Romney's artificial bounce.

Then the news stories and headlines start talking about a "resurgent Obama." The problem facing Democrats right now is that polls influence polls. When a lot of other people question their opinions, it's not unreasonable for you to question your own. This can create a bandwagon effect, where polls start to shift in a non-noisy way -- i.e., legitimately. If things even out quickly and Obama returns to his earlier standing, then count on the media to misread it all again. Which turns the tables and puts Team Romney on the defense against the bandwagon effect.

Whether any of this will happen or whether I'm just screwing around with a logical argument will become clear in a few days. Strap in and wait. This could get very interesting.

-Wisco

[screengrab source]