Another possilbility is that maybe some Democrats
are voting early because of fears of vote suppression or being shut out on election day.

Setting that aside, the AP poll went from Dem +9 to Dem +4 in partisan ID, and may have oversampled evengelicals.  But it could be that we still need to respect the GOP GOTV program.

Another thing--does partisan ID change that much within a short period?  That needs to be examined.

I'm predicting 50.2-48.4 in the popular vote with Obama winning VA (barely) and CO (by about 2-3%), and and NC too close to call.  That's 286 or 301EVs.  Ohio is just too resistant to Obama, and Florida is going to be a swing and a miss.

It's going to get closer, as the tracker average (all of them) of about 6.7% has to tighten.

Thanks for the great info, Jerome.

I'm amazed that there are hour+ lines in early voting, and the numbers are very encouraging!

Do you know how the GfK-Roper LV model compares to Zogby?  A little over a week ago, we were up by 2 in Zogby and are now up 10, where GfK-Roper seems to have the opposite (though to a lesser degree).

I'm still kind of sticking with Ras as my benchmark rather than try to overanalyze everything else.

Those polls showing a close race are bogus, period.  Nate Silver called out one of those polls and challenged the polling firm to defend its shoddy likely voter model.  

And, forget the numbers, just look at reality - one party is highly energized, extremely well-organized, and coming out in droves in early voting.  The other party is disorganized, feuding, committing gaffe after gaffe, is saddled with being the party of Bush, and has a VP candidate that is scaring independent voters silly.  

Which set of polls do you think is most consistent with political reality?

That AP poll is just impossible to reconcile with really all the other polling. I don't really understand it.

Jerome, I also think you are forgetting about Rasmussen as one of the best pollsters in '04. My sense is that Ras is pretty close to the real state of the race right now.

That AP poll is crazy on many levels...  It's obviously a Ron Fournier hack job...  He's cutting out like almost 30% of Obama supporters in his likely voter model... come on, now!

Anyways, it's good... it keeps people motivated and gets people out to vote, including those independents that we so need...

I'll stick with my prediction based on long term trends.  Obama 52, McCain 47, Other 1.

In the end former Bush voters will show up and vote against Obama.  PA is still an easy Obama win by 8-10%, Virginia will be too close to call- but so will Arizona, Florida, Missouri.

Not that it matters.  Ohio, Iowa, New Mexico, Colorado, and Nevada will make it a pretty early call.  

So basically they just threw a bunch of Obama voters out for no reason. Sure, that makes sense.

Oh, and 44% of Americans are not evangelicals.

"What if in the GE they decide they decide its a given that Obama is gonna win, and don't vote?"

That seems unlikely... In almost every important state, the race is very close:  Ohio, Florida, Colorado, Virginia, North Carolina, Missouri...  Maybe the folk in Vermont might be demotivated, but, these battleground states should remain energized up until election day.... I hope!

All things considered, it's the economy.  "The fundamental of the 04 economy was strong."  Somebody forgot to tell McCain that he was running in 08, and  not 04.  Jerome thanks for all the hard work tallying up all that data, but honestly it's a lot of wasted time.  Sorry to be a data party popper but just look at the vibes of both candidates.  McCain in his interview with Brian Williams tonight looked tired, wasted, and shook the F.U..... Obama appeared on stage with 8, count them EIGHT former prominent Foreign policy officials.... It was an impressive display.   04 was the ANTI-GAY year where there where propositions across the nation rallying up the GOP base.  This year there's nothing but A.B.O.  (anybody but Obama) rallying up the GOP.  

Finally Bush seems to be a much more stronger candidate then McCain.  Bush would have NEVER been stupid enough to pick Palin.   My God O8 should be compared more to 1980 Jerome and NOT 04.  

Period.

End of Story...

Early voting in Georgia, at least on the day I went, was not previous voters.  I saw a lot of people asking for the instructions on the Diebold machines, which have been in use for several cycles here now.  The accuracy of these things notwithstanding, they are easy to use.  Once you have used them, you shouldn't need to ask for instructions.
These all seemed like new voters to me.

I think any pollster that is relying on previous voting patterns to determine likely voters for this race, is going to end up with egg on their face.

As ProfessorReo says above and considering Silver's post earlier today, until GfK Roper tells the world how they adjusted from total respondents to likely voters, it's hard to give this poll any credibility.  

Other polls have indicated that Obama's support is more enthusiastic than McCain.  Unless there is evidence that Obama support is bifurcated between either very strong and very weak (with little in the middle of the Obama continuum), then what evidence do they have that moderate and lean Democrats will not show up.  Because he had more first time and young voters?

Plus, this is a national poll.  Half of the nation's Latino population lives in California and Texas, so the impact of a lower turnout among Latinos (which on the surface seems plausible) would be diluted in the EC count.  

For Obama, it's all about VA, CO and PA.

Thanks for the analysis, Jerome.  Can you tell me if campaigns/analysts/news media use poll results with MOEs greater than 3.5?

In the olden days, when the media had professional standards, any poll with an MOE above 3.5 was thrown out.  Same thing as I recall from advanced social statistics classes in college way back in the 70's.

Has everyone thrown out that standard these days, or has some new analysis found a means to develop reliable results using high MOE polls in the mix?

Just wondering.  Thanks in advance for your reply.

This is a great article, and I'm glad you've really delved into the internals of this poll...  I wonder if there is cause to worry about complacency...  I have a hard time believing it, but I'm a hard core voter...

I wonder if other pollsters are going to go deeper into this as well, and see if there is any complacency with democratic electorate...  I have a hard time believing it, but who knows!

Perhaps Obama needs to remind our part time democrats what would happen if McCain were to win.  Hopefully, he will do so on Oct. 29th.

Another good source of where the country is going is to look at the business channels reports on politics.  I have noticed a small but, well, noticable shift in the business reporting to a resigned belief that Obama is going to win pretty handily.  This is based mainly on comments from a variety of sources on CNBC.  As business goes, so goes the US...

I mean, look, Obama is holding friggin SUMMITS to show people how ready he is to get to work.  He LOOKS like a President...and that rubs off on voters who can be influenced.

My prediction, after the Powell endorsement, is that Obama wins big rather than a squeeker.  I know too many repubs here in MN who are either not voting or who will vote for Obama.  I know people in Ohio AND Illinois.  Very quietly, they are shaking the trees for people, to get them to vote and coincidentally, to vote for Obama.  There is a lot of, not excitement FOR Obama among voters who do not usually vote, but excitement ABOUT being a part of voting him in.

The closest thing I have seen to compare it to was New Years Eve 1999...everyone was excited to get to 2000 but there was a sliver of doubt about things still working after 12am...

Obama will be Pres, he will have a heavily leaning if not filibuster proof congress, and a reinvigorated electorate who want to roll up their sleeves and get the work that has been left undone DONE.  Like dealing with Katrina relief properly (I cannot tell you how many Dems AND Repubs want to get the proper work done, even by volunteering, and let those people get on with their lives), or dealing with Iraq, etc...

This is going to happen because Obama is the leader we HOPED he could be.

a lot to do with the early voting turnout.  Both parties I'm sure have a 72-hour GOTV operation in the works but if many Democratic voters have voted early it makes the election day efforts much easier.  As part of the Obama ground game in CO we are tasked with getting as many people to the polls as early as possible.

I'm with you Jerome regarding the blowout.  I've never seen such an organized GOTV effort as we have this year.  The rightwing blogs have this crazy idea that if we lose this year, we will be rioting in the streets.  Hell, we'll all be too exhausted to riot, win or lose. ;)

You lost me at

"(or as the low-edu commenters call "concern troll")"

Concern Troll is Concern.

:-\ sorries, maybe you should try again, without mocking your readerbase for using lingo that scares you.

Concern troll.

ps,
me must have only got ged, me am stupid :-\

Let me trot something out for a moment, and go from down the ballot on up. All indications, private and public, point to a Democratic blowout in House and Senate races. If this is the case, the pure logic of it all would tell me that Obama should win as well. Just a thought.

I just got home from a panel discussion on the election at Georgetown Law.  One of the panelists was Celinda Lake, half of the team that does the Battleground poll, and she was unequivocal in her belief that this race isn't anywhere close to a 1-point spread.  In fact, she (like every other panelist except a hedging E.J. Dionne) basically called the race as done.

The other panelists?  Byron York, Clarence Page, and Charles Krauthammer.  Everyone but Dionne called the race for Obama, and even Dionne said the odds were heavily stacked against McCain hitting the "narrow route" to victory.

The AP poll is now the top story on the ABC News web site...  They just have to TRY and make this a horse race, don't they?

they are even throwing out their OWN ABC News polling to feed the horserace narrative...  pathetic!

It is a horse race.  There are so many unpredictable elements in this campaign.  For every former GOP voter or Independent disgusted with Bush, there is another super-motivated to vote by hatred of Obama and love of Palin.  

There are probably white Dems who are having problems with voting due to issues of race.

There are probably some Clinton supporters who just are still pissed off and not voting or not voting Dem.

There are voters who just may in the end see Obama as too "foreign" "strange" and go for idiocy and comfort in the voting booth.  

I don't see that it is possible to make any assumption about outcome at the moment, just to work hard at getting the right outcome.

I do not trust Americans to make sane decisions in general elections.  I really hope I'm wrong.

And are we all forgetting that a certain someone has a certain little half-hour block of time planned for next week?

And that it was thrown out into McCain's face like a spades player throwing out high trump to flush out the opponents trump cards...worked well.  It caught McCain off guard and he has used his "best" stuff just to keep the news cycles spending all their time guessing what will be on the air-waves.  Next week will be a BIG week for Obama, just you watch.  He understands the idea of a full court press.

and a fast break.  :)

Here is an interesting [email that Ben Smith received at the Politico about voting in Evansville, IN.

I squeaked in just before the 7pm deadline to find two very frustrated poll workers and a line of a couple dozen people, due to problems with the computerized voting system not accepting people's driver's licenses. It was taking about 7-10 minutes per person just to get the computer to accept them as valid and to print out their ballot, causing very long delays.

For me the most moving moment came when the family in front of me, comprising probably 4 generations of voters (including an 18 year old girl voting for her first time and a 90-something hunched-over grandmother), got their turn to vote. When the old woman left the voting booth she made it about halfway to the door before collapsing in a nearby chair, where she began weeping uncontrollably. When we rushed over to help we realized that she wasn't in trouble at all but she had not truly believed, until she left the booth, that she would ever live long enough to cast a vote for an African-American for president. Anyone who doesn't think that African-American turnout will absolutely SHATTER every existing record is in for a very rude surprise.

There were about 20 people in front of me but remarkably not a single person left the room without voting over the 2 hours it took to get through the line.

This is one story, but I have a hunch this story is playing out all ove the country. Can Palin's devotees match this?

I think Jerome's skepticism should be welcomed, but there's a bigger wildcard out there than the Latino vote (and one which explains part how the AP got their numbers).  They assume that voters between  the ages of 18-29 will make up a smaller % of the electorate this year than they did in 2000 and 2004.  Young voter apathy makes the race competitive.  

In the previous two cycles the 18-29 vote made up 17% of the electorate (in 2000 they split evenly, in 2004 they went for Kerry 54/45).  AP's projection is that they'll make up 15% (and if you do the math they're actually dropping the majority of young voters who responded, 133 out of 254, and assuming they won't turn out).  This is almost 1/2 of the 301 voters who didn't make it through their likely voter screen, btw.

So who thinks they'll be right?  I think this is nutty.  1/2 of new registrations are from voters in this age group.  This is the group which Obama is winning 2 to 1.  Thing is, one can't say the AP is wrong because they're using historical numbers to make their projection.

A lot will turn on this.  If the youth vote turns out Obama will probably win in a landslide (and that'll have implications for the future, most of these voters, I predict, will stick with the Dems).  If AP turns out to be correct, though, we might be in for a long night (though Obama, clearly, could still win).

My hunch,btw, is that youth turnout will go up significantly, but it's anyone's guess whether they'll vote in the same numbers as those who are 20 or even 40 years older (another fudge by AP is that they're assuming 20% of the electorate will be 65+, in 2000 it was 14%).

Darcy Burner is now ahead 50% to 46% in the WA-08.

This is the sixth presidential election I've either watched or participated in. I've never seen so much intensity on the part of Democrats, not for Bill Clinton, ad especially not John Kerry. Who loved John Kerry? It was about hatred for Bush. No one is complacant this year.

It is probably worth noting here that Zogby now has this as Obama +11.9. While one can complain vigorously about Zogby (and, really, from +2 to +12 in just a few days?), one thing is pretty clear: however bad Zogby may be, he's the patron saint of good polling compared to AP/GfK. He can at least get his demographics moderately correct.

to say thanks, Jerome, for a truly incisive wrapup of where things stand today.

I agree with your analysis. We'll see what happens 12 days from now.

I'm not sure that I agree that the polls are as volatile as you say. What I see is that most polls are showing Obama somewhere between the mid to high 40s and low 50s, while McCain is somewhere between the high 30s and mid 40s. Of course, you can always take the four polls that show the most extreme spread, and call that a volatile race, but Obama sure seems to be in a much better position than either Gore or Kerry were at this point in the race. There just weren't that many polls back then that showed either of them to be consistently  ahead.