It's not as bad as they are telling us...

...but it is still not good.  The polls, that is.

RCP average has Obama up about 6.  Some fairly reputable polls have him up by 7-9.

That is not good.  BUT, considering some historical facts about pre-election polls and actual election results, it is not dire.  At least, not yet.

Over the last several elections, polls have skewed towards the democrats.  Look at exit polls in 2004 that showed Kerry winning by a large margin, for instance.  So we can reasonably predict that actual election results will be somewhere from 2-4 points more favorable for McCain than pre-election polls suggest.

Another factor is the race factor: historically, minority candidates always do better in polls than at the ballot box.  Some people just won't vote for a minority, but they would rather tell a pollster something different than expose their racism.  So we can reasonably predict another 2-4 point shift for McCain.

Together, than means a possible 4-8 point shift.  But the polls show a 6-9 point lead for Obama - and there is still a margin of error on top of that.  So Obama might be as much 10-15 points ahead in the polls (of course, he might only 1-4 points up... tricky thing, those margins of error...).

What this means is:

The race is still wide open - but Obama is still holding a slight advantage.

Now comes the last historical factor: last-minute undecideds almost always break for the known quantity.  In this case, the known quantity is McCain.  People like status quo.  Even when everyone is chanting about "change," the fact is, a lot of people would rather keep things comfortable than try something new.  This is especially true when times are turbulent.  Better to have the seasoned warhorse than they untried pony.  McCain has been going down in the polls, but I think he hit bottom and the recent Palin debate has the GOP ticket moving back up.  That means the trend will be for McCain to climb.

The questions that remain are: how high does he have to climb to ensure a victory, and will he make it?

It is not as bad as they are telling us, but it is still not good.  McCain needs to deliver in his next two debates and hammer Obama hard with targeted advertising in battleground states.  He needs to stay on offense, because Obama is much weaker defending.

This will end up a nail-biter, for sure.

Comments

[Gordon R. Durand]

OK, three more factors.

1. It's not a national referendum, it's 50 separate contests. And in the battleground states, it's a lot closer--that's why they're "battlegrounds". Sure, Obama's ahead in most, but in NV it's 49-48, in CO it's 48-45, in MO it's 50-48, in OH it's 48-46, in FL it's 50-45, and in NC it's tied 48-48. Factor in poll bias, and McCain wins them all.

2. The pollsters weight responses according to registration. The ACORN people are turning in thousands of probably bogus Democrat registrations, but the pollsters have no choice but to weight accordingly. This overstates the Obama response.

3. Most of the electoral map projections I've seen rely on a running average of the last n days of polling. It takes a while for changes to register. We won't see the full effect of the Palin-Biden debate for another week yet.

[Independent Thinker]

All good points... but...

I'm sorry to say I don't think McCain did much to help his cause tonight.

Sure, he said some good things. But he needed to say some GREAT things. Sure, Obama said some bad things. But he didn't say any OUTRAGEOUS things.

If this was football, Obama was up by a TD at the half. Tonight was the third quarter. Obama had a few turnovers, but McCain could only manage to score a field goal. He kept Obama from scoring any more, but now we are in the 4th quarter and McCain STILL needs the unanswered TD to win. Obama can give up another field goal and still win the game even without scoring again.

This is bad.

[Sid Leiken]

A reminder of a little history. In 2000 a month before the election, the polls were showing Gore 51% Bush 40%. In 1988, Dukakis 49% Bush I 41%. Gully's analysis, in my opinion, is spot on.

[Yomi Mizuhara]

I think there's only one debate left. In New York. At Hofstra U.

And McCain needs to lite into Obama, or else he's gonna have some problems.

I'll be honest, but I seriously hope Palin can do a lot more when she becomes VP.

[Michael Wilson]

Great blog and interesting views.
I think that the media had and will always try to skew the results to raise the debate.

I have found a widget that hold the poll results from different pollsters. This gives a more broad overview.

I think you might like it:-)
http://www.youcalc.com/apps/1221747067033

... and its easy to put on your blog and fits in your sidebar!

Thanks for a great blog!

Make a difference, keep on voting!

[Gullyborg]

Mayor,

I do try!

Yes, it is tight and just got a little tighter last night. McCain did well, but Obama didn't do poorly, so it is a small gain for McCain - but not enough to register much in the polls.

The polls will simply show that things are still close - but we won't really know HOW close until November 5.

That's why it is important not to give up hope of pull out of campaigning. The election will be won or lost in the next three weeks, and a lot can happen in that much time - especially in today's political environment of instant news, YouTube, etc.

[Sid Leiken]

Gully

I have had an e-mail from a person that wants me to correct a statement I made on your blog. So here it goes, on a TV report that I saw it showed a Gallup poll where Gore was leading GW Bush in October of 2000 51%-40%. I will also state that other polls were showing that their particular race was very tight.

Very similar to what we are seeing today. Gallup is showing Obama leading by 11% and a CBS poll showing Obama leading by 3%.