.

It can simply be "wrong" (bad sampling, poor methodology, intentional manipulation, whatever)... but it's almost impossible for it to be an "outlier".

"Outlier" is a statistical term... it doesn't just mean "poll is way off". In this case, the notion is that if you ran the same poll twenty times in a row, you would expect the results to fall outside of the MOE once. The problem is that this is the average of seven rolling days of polling. 2-3 of those days would have to be WAY outside of the rest to through things off this far. For a true 2-point shift, today's number needs to be 14-points worse than the number from a week ago that dropped off. That can get confused with rounding, but rounding can't account for five points in two days... it has to be a touch over 3% (and thus 11%+ difference between Mon/Tue this week and last week.

Doesn't make sense.

If someone believes there will be a eight point deficit from registered voters to likely voters there is no disabusing them of that notion.

It will get better.

Don't get it. People are stupid and easily fooled by Romney and his fucking lies. Because of them we could be stuck with a president Romney

And the fact that they overestimated the Republicans generic ballot advantage in 2010 by a whopping 250% is of no moment.

If their numbers are correct President Obama is in for the greatest popular vote drubbing since Walter Mondale. If you believe that there is nothing I can do to disabuse you of that notion.

Give it a few days see how the debate affects things...

Now he is up by two. The best he ever did in their likely voter model was tie.

. . .after this second debate has been digested and discussed by the public and the media.

Obama will be modestly ahead in the East North and West, and massively behind in the south.

That has got to be skewing this number

so it's obvious that it's an outlier. Not only that, it was posted yesterday, so you are overly concerned. The rest of us aren't, but nice try.

This is today's...

http://www.gallup.com/home.aspx

I agree it is way out there compared to the other polls that are coming out...both nationally and in the swing states...

Ohio

Their numbers are impossible to reconcile with the numbers we are seeing in other national and state polls. Either they are right or every other pollster is wrong. The law of large numbers and Gaussian distribution suggest they are.

Extreme inconsistency with other polling organizations.

over the next week.

There is absolutely no way on earth that there could be a gap of 8 points between registered voters and likely voters.

Never before has any gap remotely that big showed up in any polling.

Add to that, all the other polls which are of likely voters that either show the race essentially tied or Romney/Obama leading by a point in either direction.

Lastly, the swing state polls we get everyday, keep showing the president leading in most of them, so if that were the case then there is no way Obama could be that behind in the gallup national poll and still be leading in the swing states.

Law of averages say that when nearly every other poll shows Obama leading and gallup shows Romney lead by a massive margin then something has to be wrong with how gallup interprets likely voters.

Obama camp already exposed this.

I do think Romney's massive lead in the South is tamping Obama's lead everywhere else in some of these national polls...

Gallup Predicted Republican Generic Advantage 15%
Actual Republican Generic Advantage -6.4%

How can a the largest research organization in the world with the resources they have be so incompetent?

The Republicans intend to prevent millions of Democrats from voting.
Gallup is factoring that into their likely voter model.

It's pretty simple.

and their LV model. Plouffe has already debunked Romney's 'gains' w women and battleground states.

And Romney HAS to have OH or NV (along w/ all swings). He can't lose both. He has lost both. If 4 polls give O +4-6 for OH during Romney's peak then Romney is toast there. And he knows it. It showed last night. He did not debate like someone who has the upper hand. From the beginning he was scrambling and haggard.

But hey if folks want to cry about Gallup, talk about not facing reality...

They'll begin climbing again as Romney's good numbers drop off and Obama's take over.

If there's a flaw in their model it's unlikely it will correct itself.

After all Galup is the pollster that overestimated the Republican generic ballot advantage in 2010 by a whopping 250%.

Kos: About that Gallup poll
http://www.dailykos.com/story/2012/10/17/1145957/-About-that-Gallup-poll

Gallup is pulling people's chains.

If you look at Obama's numbers outside the South, they are comparable to 4 years ago. But in 2008 it was 50-50 in the South...today Obama is behind 61-39...

I still think Obama is going to win here in Virginia...

If one takes the internals of that poll at face value President Obama would win the Electoral College while losing the popular vote by six or seven percent. Don't forget the undecideds are not assigned.

That being said the internals and the poll itself is a joke.

bumps don't happen until about 3-5 days after an event.

So essentially you're clinging to a two week old debate that has been overwhelming replaced by two great debates where Ryan and then Romney both got their collective asses handed to them.

PLUS

http://electoral-vote.com/
277 vs 239 which means Obama would win.

From crosstabs yesterday:

Obama Romney Margin

East 52 48 O+4

Midwest 52 48 O+4

South 39 61 R+22

West 53 47 O+6

Is Obama down 22 points in the south? Per all other polls except Gallup's, no.

Look at IBD/Tipp Online's polling for the south from yesterday's poll: 44% to 51% for Romney. 7 points, not 22! Gallup is so off the mark. Are they polling KKK members in the South? If Obama was down that much in the South, FL, VA, and NC wouldn't be close. Obama would be getting killed in those states in the polling.

Also, Romney can increase his lead in certain southern states all he wants and still lose the election. But I highly doubt those southern #'s are accurate at all. Even Rasmussen (!) has Obama down 1 measly point nationally and ahead 3 points in the swing states.

they live in squalor, pretty much, and are the poorest, least educated, least healthy, least represented, and yet they vote like they're stilling living in the 1950s. Stupid as can be... they will always be poor and sick and uneducated as long as they fall for that shit, and cling to their racist, god and guns, mentality.

In Rasmussen it sure isn't. Couldn't be, as Obama is not down by more then a point in that poll. Same with Reuters, with Obama ahead by 3.

It's based on outdated models that give way more weight to conservative females in their late 50s. They don't count on the democrats who will wait 10 hours in the rain to vote at their crummy little polling place without enough machines.

And I do believe you like to shit in the pool... sorry..

much in the swing states

Polls mean nothing, wait until election day.