pageok
pageok
pageok
Obama Achieves Highest Poll Support Ever (Real Clear Politics).

Real Clear Politics computes an average of recent opinion polls.

Barack Obama has reached his highest support level ever: a 48.8% average in recent polls, all of which were conducted on Aug. 29-31. He leads McCain by an average of 4.5% (which is far from his biggest lead ever, which was about 7%).

Waldensian (mail):
Looks like they are running even in Virginia. The pundits predicted a horserace in Virginia, but to many of us who have lived here a long time, that situation is still surprising.

McCain has got to hate spending money to win Virginia. But he is going to have a hard time winning the election without the Commonwealth in his column.
9.1.2008 9:39pm
Tugh (mail):
And it doesn't even take into account the most recent USA Today/Gallup poll where Obama leads by 7 points: 50-43%
9.1.2008 9:39pm
EIDE_Interface (mail):
McCain is sunk. Might as well start getting used to "President Obama".
9.1.2008 10:13pm
VincentPaul (mail):
Yes, let's call off the election.
9.1.2008 10:16pm
taney71:
Yes, that 17 point lead that Dukakis had ensured him the 1988 election against H.W. Bush. Oh, wait it didn't. I swear people with their polls.
9.1.2008 10:25pm
cirby (mail):
Funny - a lot of pundits were predicting as much as a 15% lead for Obama by now (after a big convention bounce). Even the "pessimists" were saying he'd have 8% or so.

4.5%? That's a disaster in the making for the Dems...
9.1.2008 10:36pm
gr (www):
"Funny - a lot of pundits were predicting as much as a 15% lead for Obama by now (after a big convention bounce)."

The GOP circulated a memo predicting 15%.
9.1.2008 10:39pm
Sean O'Hara (mail) (www):
Averaging together polls with different sample sizes, margins of error, methodology, and questions gives you a number that is completely meaningless.
9.1.2008 10:43pm
DavidBernsten (mail):
All it means is that there are fewer declared undecideds now, though it would be interesting to know how many say their vote may still change by November.

(FWIW, I've been of the opinion that if Obama doesn't break 50% in the polls by the end of the Summer, he's going to lose. We'll see if I'm right.)
9.1.2008 10:43pm
gr (www):
"FWIW, I've been of the opinion that if Obama doesn't break 50% in the polls by the end of the Summer, he's going to lose."

I'm curious why this works out one way (Obama needs to break 50%) but not the other (McCain needs to break 50%) ?
9.1.2008 10:59pm
Joe Kowalski (mail):

Averaging together polls with different sample sizes, margins of error, methodology, and questions gives you a number that is completely meaningless.

This mostly true unless you know how to weight the numbers. Nate Silver over at 538 has a pretty elaborate model that weights pollster past performance, sample size, and age of the poll, then sanity checks it against state demographic data and the trendline in the national polls. All together it looks to be a good model that makes it easier to see through all the noise that you get looking at individual polls, which is quite prone to cherry picking. It will be interesting to see how well his model plays out compared to the final results.
9.1.2008 11:16pm
DavidBernstein (mail):
gr, because I think the default for the undecideds/not completely decideds will be McCain, not Obama. my hunch, based partly on the fact that Obama polled better in many primaries than he actually delivered.
9.1.2008 11:29pm
Randy R. (mail):
" my hunch, based partly on the fact that Obama polled better in many primaries than he actually delivered."

I assume you are talking about this year's primaries, the ones where Dems came out as much as ten times as republicans did. If we saw huge numbers, often record breakers, in the primaries, you can assume even bigger numbers in the regular vote. That doesn't spell well at all for McCain.

however, only time will tell....
9.1.2008 11:39pm
Smokey:
gr:
The GOP circulated a memo predicting 15%.
Can we see that memo?
9.1.2008 11:56pm
Mark Parsons (mail):

I assume you are talking about this year's primaries, the ones where Dems came out as much as ten times as republicans did. If we saw huge numbers, often record breakers, in the primaries, you can assume even bigger numbers in the regular vote. That doesn't spell well at all for McCain.


Except a similar thing happened in 1988 and that guy got a 15 point bounce out of his convention. Anybody remember the Dukakis administration?
9.2.2008 12:04am
Oren:

(FWIW, I've been of the opinion that if Obama doesn't break 50% in the polls by the end of the Summer, he's going to lose. We'll see if I'm right.)


You want to put money on that conditional bet? Set a date and a method for averaging the polls. I'll give you 3:2 on doesn't break 50 --> Obama wins up to $300 ($200 on your end, of course).
9.2.2008 12:07am
josh bornstein (mail) (www):
No one, other than the GOP and GOP supporters, ever thought that there would be a big bounce (see David B's on-point post in this thread re fewer undecided voters]. It's all a game of expectations, as we know. Will the GOP be slitting their wrists if McCain does not get a double-digit bounce from his convention? Of course not.

I just don't see any dramatic movements in the polls in this election, unless there is some dramatic external event, or where one candidate really blows it. [McCain having a stroke during a debate; Obama referring to his secret uncle, Osama bin Laden, etc.]

If the Dems can get out the vote...keep a big majority of those who came out in the primaries, then he'll win. If Palin somehow energizes her base more than we expect, or, if--in the end--people are just unwilling to vote for someone with relatively little experince [or sadly, someone who is Black], or if the Obamomentum has faded too much, then McCain will win.
9.2.2008 12:10am
Eli Rabett (www):
It's gonna be a funny old bounce unless McCain gets his act together. They already lost a day of dumping on Obama and Palin may cost them more.
9.2.2008 12:25am
theobromophile (www):
You want to put money on that conditional bet? Set a date and a method for averaging the polls. I'll give you 3:2 on doesn't break 50 --> Obama wins up to $300 ($200 on your end, of course).

I though that y'all bet beer. Wouldn't that be a six-pack to a four-pack?

The most accurate "predictor" of the winner of an election is an exit poll, as it is a survey of what has already happened, not of what people say will happen a few months down the road. Nevertheless, the exit polls in the 2004 Presidential race stated that Kerry won, 51-48, when, in fact, Bush won by a similar margin. (Here.) Likewise, as pointed out above, Obama did not poll nearly as well as exit polls indicated that he would. For some reason, it seems as if polls favour more liberal candidates, although that advantage does not materialise in the ballot box. Surely, this is another variable to correct for, which may be why DB said that Obama has to poll over 50% to have a chance at winning.
9.2.2008 4:22am
Ohio Scrivener (mail):
I am not a big fan of several of the polls referenced in the attached link. The link has three of the polls in its average relying on registered voter surveys. Likely voter surveys tend to be much more accurate. I would be very reluctant to use registered voter polls to predict the outcome of the election.
9.2.2008 4:53am
Angus:
Here is a link to the McCain campaign's 15% memo.
Link

Mark Parsons, Dukakis didn't get anywhere near a 15% convention bounce. It was right around 7%, as is normal for all political conventions.
9.2.2008 7:43am
The Ace (mail):
If we saw huge numbers, often record breakers, in the primaries, you can assume even bigger numbers in the regular vote

Historically Democrats always turn out in large numbers in the primaries.

Reference 1

Reference 2

Those 3 terms of Democratic rule since 1968 have not borne out your theory.
9.2.2008 10:59am
Hoosier:
I predict that I'm not going to argue about the polls at this point in the campaign.

The numbers for Obama have been lower that most of us thought. But this could go either way. He's running far behind the "generic Democrat," and so the undecides have reason not to support him, and so will break for McCain. OR--The trend is so heavy for Democrats this year that down-ballot votes will carry over disproportionately for Obama.

Then again, perhaps Barr will win.
9.2.2008 12:15pm
Jonathan Lund (mail):
"Yes, that 17 point lead that Dukakis had ensured him the 1988 election against H.W. Bush. Oh, wait it didn't. I swear people with their polls."

Where on Earth did you get a statistic like that? Dukakis never had a 17 point lead. Bush was beating him in the polls during this time in 1988. Dukakis was never that popular, not even with democrats. His qualities were: smart, innocuous and unlikely to cheat on his wife. I'd be willing to wager that he never even had a lead once in the polls(although after that Dan Quayle debate, he might have)but, certainly not 17 points. Obama is in good shape. He's doing a lot better than Kerry did this time four years ago. It's far from over, and it will be a close race, but no matter how you want to spin the numbers, Obama is doing very well at this point.
9.2.2008 1:13pm
Hoosier:
Jonathan Lund:

I don't vouch for Wikipedia, and they haven't sourced this. But this is what I get from their entry on "Michael Dukakis":

During this time, his 17-point lead in opinion polls completely disappeared as his lack of visibility allowed Bush to define the issues of the campaign.
9.2.2008 1:27pm
Opher Banarie (mail) (www):
No one has pointed out that it won't be the popular vote count that decides, but the Electoral College total. For an interesting (graphical) view of the state-by-state Electoral College breakdown, check out Electoral-vote.com. Hover over a state for some data, but most appear to be several weeks (or more) out of date.

They also have a variety of trend-plotting graphs that are updated every day or so here.
9.2.2008 4:05pm
Anon21:
David Bernstein:
gr, because I think the default for the undecideds/not completely decideds will be McCain, not Obama. my hunch, based partly on the fact that Obama polled better in many primaries than he actually delivered.

In point of fact, Obama did better in the primaries than the polls predicted. There were exceptions in specific states (NH being the most notable), but for the most part Obama beat his polls. The connection to his vaunted turnout operation is speculative, to the best of my knowledge, but not an unreasonable hypothesis.
9.2.2008 4:18pm
LM (mail):
I hate to say this, but I'm afraid David's prediction is right. That said, in my whole life I've predicted exactly one close election correctly -- I said if it rained in Cleveland Kerry would lose -- so I can only hope I've reverted to form.
9.2.2008 5:34pm
LM (mail):

The Ace:

If we saw huge numbers, often record breakers, in the primaries, you can assume even bigger numbers in the regular vote

Historically Democrats always turn out in large numbers in the primaries.

Reference 1

Reference 2

Those 3 terms of Democratic rule since 1968 have not borne out your theory.

Who hacked Ace's password and posted the substantive, civil comment?
9.2.2008 5:37pm
Toby:

Likewise, as pointed out above, Obama did not poll nearly as well as exit polls indicated that he would. For some reason, it seems as if polls favour more liberal candidates, although that advantage does not materialise in the ballot box. Surely, this is another variable to correct for, which may be why DB said that Obama has to poll over 50% to have a chance at winning.

I know many conservatives who do not believe in answering poll questions. Period. I know some University-town liberals who make a great point of trying to go through the line of exit polls again. Of course, I have know way of knowing if the attitudes of the people I know and observe are in any way representative...
9.2.2008 7:18pm
Aleks:
Re: Yes, that 17 point lead that Dukakis had ensured him the 1988 election against H.W. Bush.

Which was early in the summer of 1988. This late in the game a consistently maintained lead, even a small one, does start to be moderately predictive of the election.

Re: 4.5%? That's a disaster in the making for the Dems

Unless you've invented new laws of arithmetic, 4.5% is a pretty respectable lead and absent some highly improbable electoral college shenanigans, it would be a victorious margin on election day. George W. Bush won in 2004 with a smaller lead than that.


Re: If Palin somehow energizes her base more than we expect,

Energizing a base does not win elections (though it may help fill the campaign coffers). Unless people are so energized they are doing something illegal, like voting, multiple times, an "energized" vote counts the same as a dispirited one. Now let me spell it out for the Right: McCain is blowing this election by toeing the Party Line. To win this one, he needed to pull a Nicholas Sarkoszy stunt and run against George W Bush-- hard against him. Yes, he would need to keep the SoCon stuff (pro Life etc.) because he does need those voters. But he should have come up with some sort of Nixonian plan to bring "Peace with Honor" while bringing the troops home, maybe even pulled an Ike by promising to go to Iraq and make a quick end to that war. And he should be highlighting his differences with Bush on taxes, global warming etc, and promising more of the same. Maybe he should even come up with a Heritage Foundation, Romneyesque plan for universal healthcare. McCain likes to poses as a new Teddy Roosevelt, but so far he falls way short. TR was a GOP progressive. McCain is offering the same old tired ideas along with bondage to the same old special interests and power brokers. He needed to be bold and radical this year-- a lot more bold than just picking an obscure women governor for VP. The GOP can have either the White House or it can have knee-jerk allegiance to the failed ideology of the past. It cannot have both. The GOP is where the Dems were thirty some years ago: they had reached the end of FDR's legacy and could no longer run against Hoover and the Depression. The GOP is at the end of Reagan's legacy, and can no longer run against Jimmy Carter and the 60s.
9.2.2008 8:43pm