Post-debate poll

No one is fooled. Pollsters conspired to make Romney look bad. Now they’re conspiring to make him look good, so his supporters become overconfident, and drop the ball in the last month of the ground game. Tricky bastards!

The only poll that counts is the one with ballots.

MarkB, while the only poll that counts is the one with ballots, a lot of the polls being taken now are still interesting.

Eh, all the other polls–including, most notably, Rasmussen–are showing a tie or a very narrow Obama lead. So I’m skeptical of a single poll that gives such a dramatic bounce to Romney.

And I’m skeptical of all polls.

And our host should stop posting about polls because they don’t matter.

Gallup shows a tie among registered voters. Pew shows Romney leading by 4 points among likely voters. But there’s two debates and four weeks to go. We all know how quickly and drastically things can change.

Cincinnatus @ 4
I generally agree about the unreliability of polls. But polls can be useful not as a predictive tool but as a diagnostic one. If they are sampled reasonably well they can provide snapshots of a race and over time can detect momentum in one direction or another. But even then they can make no statement about the future direction of a race. They can only detect what has happened over the past few days or weeks. Polls are useful to campaigns to give them a sense of how their message is being received. They are useful to pundits in that they give them talking points to reinforce their overly inflated sense of their own importance. The real problem is that pundits and campaigns try to use the polls to create a narrative which fits into their preferred frame of reality. And this creates the possibility for a lot of self-deception.

The national polls show a tied race. As they did before the debate. Most polls have shifted a couple of points in Romney’s direction, so that they are now generally showing the race tied or a point or two in his favor, within the margin of error, instead of tied or a point or two in Obama’s favor, within the margin or error. The state polls are showing a more dramatic move in Romney’s direction, but as I have said before, they are much less reliable than national polls for showing the state of the race this early in the campaign. Their margins of error are too large to accurately measure a close race, and the pollsters are much less credible, particularly given the relatively small sample sizes.

The polls that are showing large swings, like the Pew poll, in Romney’s direction are measuring temporary euphoria. They are not predictive of the election. The last Pew poll showed, I believe, something like an 8 point Obama lead, and this one shows a 4 point Romney lead, both among allegedly likely voters. The last poll had something like a D+8 sample, and this one has an R+3 sample. That’s absurd — the party identity of likely voters in this election has not swung 11 points in the Republican direction in a week, regardless of how the debate went. What this poll tells us is that Republican overcame their normal unwillingness to talk to pollsters, because they were happier about the state of the race.

Romney will win this election — my mind has not changed on that prediction. But it’s not going to be because of that debate, or either of the next two. And it’s not going to be because the polls, today, tend to show him in a slight lead.

Steve @ 6 makes good polling points. They are useful for determining whether a race is competitive or not, and they are useful to show trends and the impact of different campaign themes. But they are not predictive, in and of themselves.

Good polls matter, bad polls don’t. The question is which is good poll and which isn’t. You really have to look at the meat and potatoes to tell which is which.

My first thought is are the increases happening in states that matter.

Steve @ 6 – excellent observations!

judge this debate 🙂

http://www.yidio.com/news/gregory-brothers-songify-best-romney-obamas-debate-premiere-7593

cheers!

The only polls that I will ‘glance’ at are polls which sample independent, swing voters. I don’t know how many of those are part of the 53%, i.e., who aren’t that 47% on the government dole 😉 but they are the ones deciding this election. I do wonder if Romney has had any movement in his direction from more libertarian type voters, especially in Virginia.