Rasmussen 9/10/08: 48-47 Obama with leaners (46-46 w/o leaners)

http://www.rasmussenreports.com/public_content/politics/election_20082/2008_presidential_election/daily_presidential_tracking_poll ^

[InterceptPoint]

What I don’t understand is how Rasmussen’s polls fail to reflect a 23 point swing in white women voters in favor of McCain. That is a pretty big share of the voting public and a big swing. For the Rasmussen polls to be showing about what they were before the conventions he must be seeing a similar swing in Obama’s direction from some other demographic group. I’d like to know what that group is.

[bushsupporter30]

I guarantee you the numbers are probably closer with independents to a 50/50 sample.

If we look at that, the 5% bounce makes sense that is showing in other polls.

Of course, am hoping the Bradley effect will counter the fraud and ground game efforts on the Rat sides so that will be a wash.

[nailspitter]

I usually think Rasmussen is the most accurate pollster, but even he admitted the other day that surging party ID for the Republicans after the convention may have rendered his turnout model, which is some months old, obsolete.

I also wonder about likely vs. registered voters in this election. My wife volunteered in Republican headaquarters the other day and a guy came in who hadn’t voted since Reagan’s last election. He’s definitely voting to keep Obama out of the WH this time. He certainly wouldn’t be included in Rasmussen’s “likely voter” pool. He simply wouldn’t be counted. but he would be counted among Gallup’s registered voters. So which poll is more accurate this time? It could be Gallup.

[Owen]

All bounces fade, but McCain’s is fading almost not at all.

Ras is now the same method he was in 2004. I believe he used previous election turnout to determine Dem/GOP/Ind target mix then. Now he uses a 3 month average, and of course 2 months ago the move back to GOP had not begun. This is the weakness of his method and he knows it. Relatively rapid changes in party affiliation won’t be captured.

Gallup is more random. It will pick up such a rapid change, but its weakness is what it has always been — that sample to sample can show large changes in party affiliation. People say “he oversampled Dems”!!!! He didn’t. He sampled randomly and that number of Dems is what he got. Big variances in that can occur and only some are real.

The result is that with the change in Ras method from 2004 you cannot compare this year’s results to 2004. All you know anything about is trend, and there is a tiny erosion of McCain bump — 1.5% max so far.

It’s all about fundraising right now. McCain’s team is fighting to get big poll bumps, even if they are unsustainable, to discourage Dem donors. The message will get far away from strong right wing quotes over the next few weeks in order to avoid energizing checkbooks.

[Crawdad]

Watch which states Obama visits. That’s all the polls I need.

[impeachedrapist]

Thanks for that data. I was just looking for it. So in August Rasmussen's gap was 9 points, now it's been adjusted to 7.6% (39.7 Ds - 32.1 Rs). Still almost two points higher than his stated identification in early September, due to the lagging rolling average.

And I have nothing but anecdotal evidence, but I'm quite certain the 5.7% gap has further closed in the past weeks since Ras' data was compiled.

[longtermmemmory]

I wonder how much operation chaos factors into the polling.

[impeachedrapist]

But at this point, only thing that worries me is the GOTV efforts and ground game as NoBama is pretty strong there and dont know if our side is as strong as 2004.

I agree, that's a definite concern for me. An article just a day or so ago was highlighting how the McCain camp is supposedly putting a lot of money into TV advertising, while the Obamessiah is supposedly focusing more on GOTV resources.

[CatOwner]

What I don’t understand is how Rasmussen’s polls fail to reflect a 23 point swing in white women voters in favor of McCain. That is a pretty big share of the voting public and a big swing. For the Rasmussen polls to be showing about what they were before the conventions he must be seeing a similar swing in Obama’s direction from some other demographic group. I’d like to know what that group is.

Exactly.

[MrB]

Polls historically have given the dhimmirats a few points more than the actual election results.

0bama will have to be 5 points up to make it an actual elective close race.

[scooby321]

Zogby is the best Pollster. He was the only one who got GW DUI loss in the polls.

The actions of Obama is the only Poll I need to see.

I see the crowds and Obama has lost appeal and Palin has them overflowing.

[hoe_cake]

“So if you think that Dems will outvote Pubbies by about 9 points in November,”

Well, then, let’s hope the young Obama crowd has smoked too much the night before to limp over to the polls.

[bushsupporter30]

We will find out in the next two weeks of the states that the campaigns are going too and hopefully will give the McCain campaign and the RNC time to get GOTV going in those states.

But I feel confident with the enthusiasm level. Thats what got us the WH in 2004 and I think the level now definitely exceeds enthusiasm levels in 00,02 and 04.

[KeatsforFirstDog]

I understand his 3 month rolling party affliation, but it would be nice if once or twice a week he’d just give the bare bones numbers without any tinkering, you know?

I suppose I could mathematically do the numbers if I get time later today...

[pgkdan]

There's a problem with Rasmussen's party id percentages. He weights his poll according to national averages and as a result samples a lot more democrats than republicans. But from what I heard on the radio this am he may want to adjust his sampling. There's a poll out this week...can't tell you which...that has party id among those who claim a party affiliation at 48% dem and 47% GOP. They claim they've seem a huge swing since the GOP convention.

Rasmussen said that if there was a sudden swing in party affiliation percentages that his polling would be way off until he caught the swing.

[se_ohio_young_conservative]

I see it this way. In the Miracle on Ice. The Soviets did not get blown out. What stunned them is when it was close going into the 3rd period. They were stunned and never got it together. Expecting a blow out and getting a hard fight.

We should not even be in this race right now. The Democrats should be up 10 or 15 points. CONSIDER WHERE WE WERE 6 MONTHS AGO.

[Blood of Tyrants]

Pay no attention to national polls. It is the state polls that count and in Ohio, Michigan, Pennsylvania, Wisconsin, Virginia, North Carolina and Florida are showing a distinct McCain/Palin shift.

[impeachedrapist]

I hadn’t yet heard of the 48-47 numbers, but I can believe them. The rest of your post is exactly what I was saying - or trying to say :-) - up above.

[pgkdan]

I wish I had caught the name of the polling outfit but I was driving into work and missed it but was knocked out by the numbers. McCain could really be up 8 to 10 points at this point.

I remember 1980 and how close everyone thought the race was...right up until Carter conceded at 8:00 PM EST.

[libscum]

Hey Rasmussen, if Mccain is winning the majority of white women, indys and white males just tell how the hell is this thing stil TIED? He needs to look at his sampling techniques.