Home › Blogs › Sean Oxendine's blog ›

Ugh

Submitted by Sean Oxendine on Tue, 11/04/2008 - 07:36

Before Sunday, McCain was actually in a better poll position than many imagined.  Not that he was in a good poll position, but it wasn't hopeless.  The tracking poll range was about a 2 to 9 point lead for Obama, which was actually the final ranges for Bush against Gore.  As we know Gore ended up winning the popular vote, and about half the country is convinced he won the electoral vote as well.  There was also a good case for McCain getting the lion's share of the undecideds, given their demographic makeup and the fact that Obama had spent three quarters of a billion dollars and had still not won them over.  Add into that the fact that the difference between the 2-point poll and the 9-point poll appeared to be structural, having more to do with the model for likely voter turnout than simply arising from random variance, and you had at least a straight-faced argument for McCain winning.

Sunday night, that changed.  All the the tracking polls moved toward Obama.  Even IBD and Battleground, the two polls whose models seemed to be the most favorable toward McCain, shifted from 2-point races to 4- or 5-point races.  At that point, even under the best-case McCain turnout scenario, it became much harder to argue that undecideds and error margins could result in a McCain electoral win.

At that point the only hope was that this would be a one-day blip.  That's part of the reason that I've held off doing electoral college projections -- I wanted to see what the trackers did today.

I'll tell you, it ain't pretty.  Zogby has gone from an 8-point Obama lead to a 13.5-point Obama lead.  IBD/TIPP went from O+4.5% to O+7%.  Given that tracking polls are rolling averages, and are therefore "sticky," the move is probably even more pronounced than we are seeing.  In other words, undecideds seem to be breaking heavily for Obama.

Obviously when your candidate is down two to nine points to begin with, this is not a good thing.  At this point, the polls would have to be catastrophically wrong for Obama to lose.  Polls have been substantially wrong before.  They might be now.  If you want to help make them wrong, you have to get out and vote!

5
Your rating: None Average: 5 (1 vote)

Comments

RealClearPolitics'04 polling averages vs. actual outcome

Submitted by mlindroo on Tue, 11/04/2008 - 07:56.

> At this point, the polls would have to be catastrophically wrong for Obama to lose. 

> Polls have been substantially wrong before.  They might be now.

 

True -- *some* polls are wrong. But in this particular case, ALL of them (even including the ones making the most conservative assumptions re. first time or minority voters) would have to be disastrously wrong!

For what it's worth, here is a table summarizing the last RealClearPolitics polling averages versus the actual outcome four years ago.
While it's true that the pollsters significantly "misunderestimated" GW Bush's support in Florida and Missouri, overall the battleground state polls ended up being slightly overly pessimistic about Kerry's prospects. In the end Kerry actually fared 0.56% better than expected!

 


  Bush Projected Bush actual
CO 5,20 % 4,70 %
FL 0,60 % 5,03 %
IA 0,30 % 0,67 %
MI -3,50 % -3,43 %
MN -3,20 % -3,50 %
MO 4,20 % 7,24 %
NH -1 % -1,37 %
NM 1,40 % 0,80 %
NV 6,30 % 2,49 %
OH 2,10 % 2,12 %
PA -0,90 % -2,52 %
WI 0,90 % -0,38 %
 
Considering that Obama certainly is not more liberal or more unappealing/a worse campaigner than Kerry was, and the comparative lack of enthusiasm for McCain compared to Bush'2004 among GOP partisans, I would be very surprised if there is a motherlode of hidden McCain votes in this year's data... In particular, I note the seemingly solid Obama lead in formerly red NV, NM, CO, IA, PA and VA where not a single pollster thinks McCain is within 3 points (the average is > 5% +). NH is also in this category. Obama wins 291-247 even in this case,  where McCain gets all the lucky breaks and wins every single race where he is within striking distance!
 
We will know how this plays out within the next 24 hours, but in the light of Kerry's relative strength it would difficult to attribute a stunning Obama loss to anything except some weird race-related dynamic that does not show up at all in the polls. This would undoubtedly be a very ugly outcome considering also that McCain certainly will lose the popular vote no matter what happens.
 
MARCU$

 

More Likely Scenario

Submitted by Sean Oxendine on Tue, 11/04/2008 - 08:02.

I actually didn't have American politics in mind when I noted that polls had been catastrophically wrong before.

http://en.wikipedia.org/wiki/Shy_Tory_Factor

But like I said, this is an extreme long-shot.

Most pollsters, in the week

Submitted by williamsjoseph on Tue, 11/04/2008 - 08:02.

Most pollsters, in the week before the election, want to be as accurate as possible. In this way they can get new clients for the next cycle. With every poll showing McCain behind, it would take a tsunami of a Bradley effect for McCain to win. America seems to want change, by careful what you wish for.

well, this guy ain't buying it

Submitted by Ironman on Tue, 11/04/2008 - 08:18.

he may be all wet, but unlike some conservatives , he's not throwing in the towel before the damm polls close.

http://seanmalstrom.wordpress.com/2008/11/03/toast/

Let's see; we are only dealing with a hostile press, an opponent with a $600M warchest, a stock market plunge and a president who is less popular than dandruff. And with the exception of Paulsen's Folly, we knew all this going in, now didn't we?

To paraphrase Laura Ingraham--shut up and VOTE!

I'm counting on something akin to...

Submitted by Darvin Dowdy on Tue, 11/04/2008 - 08:35.

...the Bradley phenom.  Nothing to do w/race, but I'm hoping when those 8% undecided voters step up to that voting booth, they have an epiphany of sorts.  Realize they simply can't turn the nation over to some inexperienced socialist.  And then pull for McCain/Palin. 

I would like to say it has been a real pleasure bantering back in forth w/all the good folks at this site over the past weeks/months. And a big thanks to those enlightened folks that post their insightful essays which really put things in perspective.   I'd say its time to take a few days off from politics.   After which it'll be time for us Movement Conservatives to wrest control of  "our" GOP from the scoundrels and rebuild. No small task.  Not even sure it can be done. DD

Remember the day before the NH primary?

Submitted by Ironman on Tue, 11/04/2008 - 08:43.

Poll: Obama opens double-digit lead over Clinton

 

http://www.cnn.com/2008/POLITICS/01/06/nh.poll/index.html

I think (no: hope!)  the NH

Submitted by mlindroo on Tue, 11/04/2008 - 09:03.

I think (no: hope!)  the NH primary is largely irrelevant due to the comparatively small difference between the two Dem candidates (O and Hillary). Realistically, current Obama voters will be less inclined to switch to McCain than they were in the primaries, when the main question was who would be most competitive in the general election.

 

MARCU$

IBD Numbers

Submitted by jmumich on Tue, 11/04/2008 - 08:48.

Check out the IBD numbers here:   http://www.ibdeditorials.com/series13.aspx?src=POLLTOPN.

The actual final tracking poll numbers appear to be O 47.5% M 42.4%.  IBD then allocates undecideds, adding 4.0% to Obama, and 1.9% to McCain.  In other words, they expect undecideds to break roughly 67% Obama and 33% McCain.  

I'm not sure what their methodology was for doing that, but the actual result could just as easily put it the other way around, adding 4 to McCain and 1.9 to Obama puts it at O 49.4% to M 46.4%.  Still a likely Obama popular victory, but within the margin of error and close enough to allow McCain to squeak out an electoral victory by picking off PA.  

Of course the third party candidates won't be getting 4.2% either.  

I'd be wary of these final numbers, since it seems that pollsters are apt to rather arbitrarily allocate the undecided voters.  

Kick back with a Bud

Submitted by cr on Tue, 11/04/2008 - 09:15.

It's (not so) funny watching people trying to wring some hope from the polls. Today is the day to sit back and let it go.

Whatever is coming is already on it's way.

Market says Gore won Florida

Submitted by RisingTide on Tue, 11/04/2008 - 09:38.

... people bet on elections, ya know?

My prediction: Obama w/52% of popular vote

Submitted by clam75 on Tue, 11/04/2008 - 09:54.

Sean, I think that Zogby's a bit high this morning.  Rasmussen (and I think their partisan weighting is a fairly good method) has a 52/46/2 race, and I think that will prove to be largely accurate.  For us to have a 1952-in-reverse blowout (Ike got 55% of the vote & won in TX, VA, and TN), I think Obama would've had to reach 52% at the beginning of October and stay there.

BTW, I have enjoyed your metric insights quite a bit this cycle.  A blogger a few weeks back compared you to Jay Cost, and I agree every bit.  While this cycle hasn't turned out as I'd hoped, you've certainly been a very keen observer.

300

Submitted by Evil Conservative on Tue, 11/04/2008 - 11:15.

We are with you, McCain!

For America!

For FREEDOM!

To the Death!

polls

Submitted by DMcCain on Tue, 11/04/2008 - 14:23.

In a few hours we'll know which polls have been able to accurately forecast this historical election. I for one hope these polls are accurate.