Pollster.com

Articles and Analysis

 

Morning Status Report for 10/13

Topics: Status Update

Sunday was, as usual. a light day for new poll releases other than the national daily tracking polls. Since our last update Saturday afternoon, the only new statewide polls we had logged in battleground states were new surveys from PPP in Colorado and the new University of Cincinnati poll sponsored by a consortium of Ohio newspapers. The two news polls are counter trend in opposite directions. new Colorado poll shows a bigger Obama margin than other recent polls, while the Ohio surveys shows McCain with a two point advantage.

2008-10-13 Daily.jpg

The national daily tracking polls do, for the most part, show slight shifts back to Obama since the middle of last week, although Obama's margin on our national trend estimate as of this writing is almost exactly where it was this time last week.


2008-10-13 trends.jpg

 

Comments
JFactor:

"The national daily tracking polls do, for the most part, show slight shifts back to Obama since the middle of last week, although Obama's margin on our national trend estimate as of this writing is almost exactly where it was this time last week."

That is a key phrase that most people forget when immersed into the world of daily tracking polls. The trends are not in McCain's favour.
____________________________________________
http://www.internationalpoliticstoday.com

____________________

vaview:

I'm not completely familiar with your poll-screening on this site, but I thought you included nearly every poll. I just wondered why there is a new North Dakota poll (showing an Obama lead) over at electoral-vote and it is not listed here. I'm certainly not suggesting bias, but is there a problem with that poll that electoral-vote isn't picking up?

____________________

MasterFramer:

yeah if you are going to include the Ohio outlier then the ND poll should be on the site as well

____________________

bythesea:

Can McCain still pull it out? Yes of course its possible. The thing the Obama campaign has to watch out for is a repeat of its behavior in the primaries. Once the inevitability factor came into play Obama employed a prevent defense strategy. And every football fan knows taht can be a disastrous decision. Clinton won lots of the later primaries as Obama just sat on his lead and didn't fight. The same practice now could lead to disaster. McCain won't pull his punches as HRC did. If he starts connecting on those little dink passes, baby watch out.

____________________

st paul sage:

I'm a long time pollster reader and I think you and Nate Silver are the best analysts of the polls out there. I've been talking a lot to my nervous democratic friends about the "Bradley/Wilder Effect" and I can't seem to draw on the right data. My answer tends to be 2 points:

1) There is simply not enough data to project much of anything from past statewide general elections to this year's national election.

2) But what data there is suggests that there is no such effect. While Bradley and Wilder may have underperformed the pre-election polls, Harold Ford '06, Deval Patrick '06, Barack Obama '04, and Carol Mosely Braun '92 did not and since they are more recent, they are more reliable. But I vaguely remember that Ford probably outperformed the pre-election polls and I don’t have any data in front of me. You do. So am I right? How did Ford, Patrick, Obama, and Mosely-Braun do on their election days vis-à-vis the pre-election polls? And if you already covered this in a post, I apologize for missing it.

Furthermore, Lance Tarrance, who was Deukmejian's pollster against Bradley in 1982 just posted a great article at RCP basically stating that according to his polling, THERE WAS NO BRADLEY EFFECT IN 1982 and in fact, Deukmejian was closing on Bradley in 1982 steadily and his final result matched the polls, pre-election. I don't think primary elections are even vaguely similar to general elections so I don't think it's worth talking about them.

____________________

Ulysses:

I am an ex-McCain supporter, and new Obama supporter in Pennsylvania. So, I guess you could say that I'm indicative of a trend that I see among many of my friends. And probably is best summed up -- in terms of the 'why' of this election -- by this inspiring, young lady:

http://news.bbc.co.uk/2/hi/americas/7665925.stm

____________________

Luiz:

Mark,

Zobgy seems an outlier to me. I wonder if you could comment on that.

____________________

MasterFramer:

@ Luiz

It has been all year

____________________

JFactor:

I think Zogby uses a party ID model that has Dems +2 over Republicans which is of course ridiculous. I'm not sure though if they got rid of sampling by party ID completely recently, can anybody confirm?
_________________________________________
http://www.internationalpoliticstoday.com

____________________

Luiz:

@MasterFramer
@JFactor
Thanks

____________________

falcon79:

@ Mark Blumenthal:
Sir, I am wondering why Virginia has not been painted blue on the map?
Some time back, you said that a state was painted a toss-up if it was 3.7 points or less in favor of either candidate.
As of today, Obama holds a lead of 3.8 points, which is just enough to cross the threshhold to make the state blue.
If a cut-off is treated with flexibility, does that not debar the cutoff from being a cutoff?

____________________

brambster:

"Furthermore, Lance Tarrance, who was Deukmejian's pollster against Bradley in 1982 just posted a great article at RCP basically stating that according to his polling, THERE WAS NO BRADLEY EFFECT IN 1982"

DUHHHH!

Seriously. Does anyone really think that 5% of voters walked into the polling booth, voted one way, then walked out and told pollsters that they voted for the black candidate when they did not? That's also a net of 5% when you subtract those that voted for the black candidate and then lied or otherwise screwed up and said otherwise.

This is such a big number that it's perposterous to think it was ever true. Every other effect was likely the result of bad polling, or so small that it was imperceptible.

Find me a person that claims to have been so ashamed about not voting for a black candidate that s/he lied to a pollster in order to save face by claiming that they voted for a black candidate.

____________________

brambster:

@falcon79

States are ranked by their 65% and 95% confidence intervals, and these intervals are different for each state based on the number of polls and the variability of the results, and this changes on a daily basis. Pollster is careful about not adjusting their model mid-stream in order to help it fit some pre-determined target. It is what it is and it will turn blue when it breaks the 95% confidence interval.

____________________

falcon79:

@brambster:
thank you for taking the time to explain this to me.
i did not realize that there was a 95% CI involved, and really did think that Mr Blumenthal referenced a point-estimate cut-off of 3.7% some time earlier this month.

____________________

Pachete:

St Paul Sage-

While i dont believe the bradley effect will be a large or deciding factor in the GE and ive read alot of things suggesting the bradley effect disappeared in the mid-late 90s, we have to keep in mind that its never been tested on the national level before so it may be accented in some states.

____________________



Post a comment




Please be patient while your comment posts - sometimes it takes a minute or two. To check your comment, please wait 60 seconds and click your browser's refresh button. Note that comments with three or more hyperlinks will be held for approval.

MAP - US, AL, AK, AZ, AR, CA, CO, CT, DE, FL, GA, HI, ID, IL, IN, IA, KS, KY, LA, ME, MD, MA, MI, MN, MS, MO, MT, NE, NV, NH, NJ, NM, NY, NC, ND, OH, OK, OR, PA, RI, SC, SD, TN, TX, UT, VT, VA, WA, WV, WI, WY, PR