Poll: Obama leads in 3 of 4 key Bush counties

http://www.politico.com/news/stories/1008/14556.html ^

[tomymind]

Barack Obama has erased traditional Republican advantages in four key bellwether counties that President George W. Bush won in 2000 and 2004, according to a new Politico/InsiderAdvantage survey. Each county is critical to the outcome in the battleground state where it is located.

In Reno, Nevada's Washoe County, Obama leads McCain by a 46-45 percent margin, with six percent undecided. Obama posts a wider 50-44 percent lead with five percent undecided in Raleigh, North Carolina's Wake County, and another 6-point lead in Hillsborough County, Fla., where Tampa is located. There, he edges McCain 47-41 percent, with 11 percent undecided.

Among the four counties tested, McCain leads in only one: Jefferson County, Colo., a populous Denver suburb. McCain is ahead there by a margin of 45-43 percent, with eight percent undecided.

At first glance, these Politico/InsiderAdvantage numbers might not look so troubling for McCain, who trailed Obama by 10 points in an ABC/Washington Post national survey, released Monday.

But these four counties are crucial battlegrounds in four of the most competitive states in the presidential race. In recent years, the Republican path to the White House has run through these areas.

In 2004, President Bush won Washoe County, Nevada's second-most populous county, by a four-point margin over Massachusetts Sen. John F. Kerry. This year, when Obama is expected to run up a big vote lead in Las Vegas' Clark County, McCain is unlikely to be able to afford a loss in Washoe.

[Irish Eyes]

I don’t think so: I just think they are all operating from an “inside the beltway” mentality-—even if they aren’t located there-—and they are believing this nonsense about a 10% Dem “likely voter” advantage, which has never, ever proven to be even remotely close historically. In 2004, after all the Howard Dean internet “revolution” and all the anti-war, Move-On stuff, the breakdown was 39%/39% Dem GOP.

I think that you are right. The elephant ( ?.. lol) in the room is the fact that Obama's candidacy has brought a tremendous amount of enthusiasm to the black community. Many in this community think that is is their time and nothing else makes any difference. Howard Stern did an interesting experiment about the lack of knowledge of the man on the street in respect to Obama. I do not think that the pollsters know how to factor in this phenomenon.

[Perdogg]

Wake Co a key county? - please!!!! Bush won Wake by only 2% in 2004. Actually Kerry did much better in Wake than Gore did. As long as McCain talks free markets he should well.

When McCain visited NC he visited Wilmington, which is a bit odd. If I am “losing” in NC I wouldn’t waste time in Wilmington. I would hit

[princeofdarkness]

Sorry, I thought you were referring to strictly voter election results.

[moose2004]

zzzzzzzzzzzzzzzzzzzz...........

[LS]

I deal in people. I can't think of one person---and I know a lot of people--who has changed party registration in the past 8 years except . . . my wife, who changed in the primaries to Dem to vote against Hillary.

Hmmm, suppose she'll be voting for Obama?

Now, here's some "inside baseball." I've done precinct work for the Republicans for 10 years. We have voter ID and voting RECORDS going back for five elections. We can tell who is a "strong Republican," not so strong, and quasi-independent. When our walkers go around, they make notes. These notes get back in the system.

Very, very few Republicans change party ID. Even those who hate/hated Bush have not changed party ID. We have very high confidence that our voters are "our" voters. I have talked to them on the phone, election after election. I have never once heard, "I used to be a Republican, but I switched." We have the indies marked, too, and I've had them answer both ways, "I switched from Republican/Democrat."

You still gotta do better---a lot better than Blumenthal.

[LadyNavyVet]

“I deal in people.”

IOW, don’t give me facts, I’d rather poll my circle of acquaintances? LOL!

I’ve worked for the party here in FL, too and walked precincts. I know a thing or two about GOTV. And this time so do the Dems, who took the Pubbie’s GOTV methodology, coopted it, and if what I’m seeing is any indication, improved on it. Obama has over 50 field offices and 150,000 volunteers here. I live in the swing part of Florida, and they’re everywhere. Kerry had nothing like it, and McCain better get off his dead butt, since voting has already started.

And while we’re trading anecdotes to avoid talking about facts, I know people who voted for Bush in 2004 who are voting for Obama. Security moms, who voted for Clinton, who are going back to the Dem party this time. No amount of persuasion works with them. I’ve tried.

Now, back to facts. Voter ID is not a static idea, and when voters answer the question on a poll they are often not taking into consideration their party registration. Also, some states don’t require voters to register with a party, and in recent years the trend has been toward eschewing both parties and claiming independent status.

Here are some facts from the 2004 NAES (yep, those pesky facts again!)

http://www.ou.edu/policom/1501_2005_winter/roundtable_kenski.htm

“You still gotta do better-—a lot better than Blumenthal.”

Blumenthal was talking about the NAES. Based on over 81,000 interviews, party ID was highly labile in 2004, and ranged between +7 for the Dems to +2 for the Reps. It was about +4 for the Dems on election day, and tanked shortly afterward, as respondents no longer wanted to be associated with the losers.

For the third time, do you have any facts to back up your assumptions? Little stories about your wife are cute but unconvincing, especially since you present yourself as guy who knows about polling, yet you shy away from discussing it in post after post after post. Hmmmm....

[LS]

I gave you the facts. You just ignored them. I explained that we have our own voter ID base, and that it is phenomenally accurate. Voter ID is pretty static.

[LadyNavyVet]

LOL! I give you data you give me voter rolls. Dems have voter rolls, too.

NAES isn’t he only pollster who has found party ID to be highly variable. Ras has found the same thing, and he interviews 500 every night.

http://www.rasmussenreports.com/public_content/politics/mood_of_america/party_affiliation/party_affiliation/summary_of_party_affiliation

[LS]

LOL. (and big sigh). Voter rolls are data last I looked. You’re being circular, and it’s getting silly. If the GOP has on its rolls 100,000 Republicans in Montgomery Co. (out of, say, 250,000) and if exit polls confirm that 100,000 people voted for Bush, it’s a darn good likelihood that those people were the Rs. And we know the exact party breakdown of our districts, including the percentages of registered voters.

[Thane_Banquo]

Problem is, you're assuming Ras polling surveys, and other polls, reflect party ID reality, rather than final results. Statistically, if you have pre-election polls that show hugh variation in party ID, and exit polls over the past 20-30 years that show party ID remarkably stable, then the stable data is more likely to be accurate. This is simply probability. So what's going on with Ras surveys? Hang-up rates, among other things. When a given side is doing well in the press, there is going to be a lower tendency to hang-up on pollsters relative to the opposite situation. I believe that because of the media establishment's hugh support for Zero (more than any other candidate before), McCain supporters are fed up with the media establishment, and are hanging up in higher numbers. Moreover, it has become politically incorrect in our culture to admit to opposing Zero, because the media has for weeks months been on the drumbeat of "you're a racist if you don't vote for 0bama." When I was polled last night, for instance, even I felt a bit shy about admitting that I had a very unfavorable opinion of Zero and that I was voting for Sen. McCain. It was a weird moment for me. These kinds of things affect hang-up rates, and the number of people who punt by saying they are "undecided" when they really aren't, but there is no reason to expect them to affect actual votes cast in private. I mean, 15% undecideds 3 weeks before an election like this is simply silly on its face! There is, I believe, a very big 0bama effect going on, which is composed partly of an underlying Bradley effect, and an additional MSM thought-gestapo effect, where opponents of Zero are marginalized and thus silenced. We can't expect polls not to be influenced by this. The alternative explanation you seem to be positing, if I understand it correctly, is that people call themselvse Democrats one week, then go through some great existential awakening and start calling themselves Republicans the next week, only to backslide and start calling themselves Democrats again the week after. Lather. Rinse. Repeat. Choosing to side with a party is a deliberate act, often requiring that you go to an eleciton office and register with that party. It represents a deliberate decision that one party has a better agenda for America. There is no reason to believe that people would swing back and forth over that. So what is my evidence that these polls are oversampling Dems, and not getting true unbiased samples? Newsweek shows McCain under-polling Zero among senior citizens. Some polls show McCain losing white men by double digit margins, for goodness sake! That simply doesn't pass the smell test.

[Thane_Banquo]

This time with formatting:

Problem is, you're assuming Ras polling surveys, and other polls, reflect party ID reality, rather than final results. Statistically, if you have pre-election polls that show hugh variation in party ID, and exit polls over the past 20-30 years that show party ID remarkably stable, then the stable data is more likely to be accurate. This is simply probability.

So what's going on with Ras surveys? Hang-up rates, among other things. When a given side is doing well in the press, there is going to be a lower tendency to hang-up on pollsters relative to the opposite situation. I believe that because of the media establishment's hugh support for Zero (more than any other candidate before), McCain supporters are fed up with the media establishment, and are hanging up in higher numbers. Moreover, it has become politically incorrect in our culture to admit to opposing Zero, because the media has for weeks months been on the drumbeat of "you're a racist if you don't vote for 0bama." When I was polled last night, for instance, even I felt a bit shy about admitting that I had a very unfavorable opinion of Zero and that I was voting for Sen. McCain. It was a weird moment for me.

These kinds of things affect hang-up rates, and the number of people who punt by saying they are "undecided" when they really aren't, but there is no reason to expect them to affect actual votes cast in private. I mean, 15% undecideds 3 weeks before an election like this is simply silly on its face!

There is, I believe, a very big 0bama effect going on, which is composed partly of an underlying Bradley effect, and an additional MSM thought-gestapo effect, where opponents of Zero are marginalized and thus silenced. We can't expect polls not to be influenced by this.

The alternative explanation you seem to be positing, if I understand it correctly, is that people call themselvse Democrats one week, then go through some great existential awakening and start calling themselves Republicans the next week, only to backslide and start calling themselves Democrats again the week after. Lather. Rinse. Repeat.

Choosing to side with a party is a deliberate act, often requiring that you go to an eleciton office and register with that party. It represents a deliberate decision that one party has a better agenda for America. There is no reason to believe that people would swing back and forth over that.

So what is my evidence that these polls are oversampling Dems, and not getting true unbiased samples? Newsweek shows McCain under-polling Zero among senior citizens. Some polls show McCain losing white men by double digit margins, for goodness sake! That simply doesn't pass the smell test.

[LadyNavyVet]

You continue to confuse party registration with party ID in polling. They are not the same thing at all, as anyone who really knows polling, as opposed to internet poseurs, will attest. I have been very careful to use “party ID,” not “party registration” in my posts, since they are NOT the same thing.

Voter registration rarely changes, since it’s a pain to go to the Supervisor of Elections and reregister just because right now you like a different party better. Party ID in polling, however, changes in response to news events. It is labile, not static. That is hard for ideologues to understand, since their party ID and party registration are in lockstep. But to those not ideologically wedded to a party, and that is a greater and greater share of the electorate, party registration may not change because that requires effort, but party ID changes frequently in response to events in the news. That has been replicated in polling time and time again over many years.

And no, just because there are a certain number of Rs in Montgomery County and that matches the vote count for the Pubbie, one cannot say that THOSE voters are the exact ones who voted for the Pubbie. There are crossover voters in every election, and right now, according to Gallup, Obama is winning 5% of conservative Republicans and McCain is winning 3% of liberal Democrats.

[LS]

First, I've been polite in not calling you names, so please don't start with your little "poseur" stuff. I've done enough statistics, both in politics and in education, to know what you are talking about.

Second, if you're going to play the "separate the party ID from voter registration" game, then we'll take that a step further: every single "Battleground Poll" contains the question on conservatism/liberalism, and every single time the LARGE majority of Americans consider themselves more or somewhat more conservative. Only 8% consider themselves liberal. So that's a darn good plug-in for party ID.

But we're done. You were burned in 2006 and now want to impress everyone with how much you've learned since then. We're impressed. And I think you're wrong, just the same. But we'll find out in a few weeks. However, if the internal stuff I know about the metrics here in OH is accurate, Obama is toast, and the polls are wrong, just as some of the other posters say---and for exactly the reasons they say.

[LadyNavyVet]

“First, I’ve been polite in not calling you names...”

And I was too, until your “Actually, if you know anything about the ‘insides’ of campaigns” crack. I’m always polite as long the other poster is.

I agree with you about OH. But Obama doesn’t need OH. McCain does, just like he needs FL, VA, CO, NV, etc. All the current battleground states are ones that were Bush country in ‘04. Even if this race tightens up, and I think it will, McCain has an electoral mountain to climb.

[LadyNavyVet]

“Statistically, if you have pre-election polls that show hugh variation in party ID, and exit polls over the past 20-30 years that show party ID remarkably stable...”

Exit polls over the last 20-30 years have not shown party ID to be remarkably stable. In 1980 it was +15, for instance.

“McCain supporters are fed up with the media establishment, and are hanging up in higher numbers.”

We hope, we don’t know. That’s called the “Shy Tory” effect in polling, BTW. These polls that have Obama up right now had Bush up in 2004. Are people more fed up with the media now and not answering pollsters than in 2004? We don’t know. Pollsters will tell you that the hardest demographics to poll are the youth and minority vote. The oldsters and whites are by far the easiest.

“I mean, 15% undecideds 3 weeks before an election like this is simply silly on its face!”

Yes, it is, but Freepers are hanging on that TIPP poll, which has the highest rate of undecideds. Most pollsters have undecideds down to 5% or so, about right by historical standards.

“Choosing to side with a party is a deliberate act, often requiring that you go to an eleciton office and register with that party.”

There is a big difference betweeen party registration and party ID in polling. Most people don’t bother to change their registration every time they decide they like the other party better. Polls ask what people CONSIDER themselves to be, not what their registration is.

“Newsweek shows McCain under-polling Zero among senior citizens.”

They’re getting their 401K statements, and they need that money to live on right now. It’s always “the economy stupid” that is the greatest drive in voting behavior.

[Thane_Banquo]

Second, the Reagan administration brought a hugh secular change in political ID, mostly as southern Democrats became southern Republicans, and the GOP became generally became a more palatable party. Note that the party ID changed dramatically from before 1984 to 1984 and after. The idea that a very controversial figure like BHO is going to bring a lot more Americans into the Dem Party before he has even served a term in the WH doesn't really pass the smell test. Even Reagan required 4 years in the WH to change people's attitudes towards the GOP. And as I'm sure you know, the every day regular Americans absolutely loved Ronnie. The same cannot be said for BHO. As for BHO's turnout rhetoric, let's be honest: Dems have been registering black voters every election for decades, and it seems to have no effect. Do we really believe there are a whole bunch of eligible black voters who have never voted and now intend to? How do we know that these "new registrations" aren't merely re-registering people who were already registered?

[xsmommy]

has it occurred to you the Mccain camp may want to allow the other side to feel over-confident? it's kind of the corollary of our side feeling defeated. so yeah, there really IS a reason to have our side saying that when it isn't actually true.

[tatown]

“zzzzzzzzzzzzzzzzzzzz...........”

.....................

Without question the most prophetic thing I’ve ever seen come off your keyboard. Nice job...

[LadyNavyVet]

“Dems have been registering black voters every election for decades, and it seems to have no effect. Do we really believe there are a whole bunch of eligible black voters who have never voted and now intend to?”

This time, yes. I’ll believe in the huge youth turnout the Dems are predicting when I see it, but blacks will crawl on cut glass for “Barack.” Reports are that 37% of early voters in GA are black, even though they only make up 29% of the population. And Survey USA, in it’s state polling, is reporting early or absentee voters. Big numbers for BO. (One caveat—the SUSA numbers are, by their nature, small sample sizes.) Let’s hope they’re black, ‘cause if they’re youth, the good guys are in trouble.

[Mr. Silverback]

There is no advantage for McCain to tell the world he’s down, yet he did it anyway.

OK...so every poll had him down by margins outside the margin of error and a couple of widely reported national polls have him down by double digits you think it makes him look bad to mention that he's down in the polls?

Huh?