One of the things we're doing with our polls right now is asking folks who they voted for in 2004, so that we can get as a clear picture as possible of what voters are changing their preferences and fueling this large movement toward Obama.
This goes against the conventional wisdom a little bit but some of Obama's strongest gains in Ohio have come among small town voters. Folks in those places reported voting for Bush by a 54-35 margin but are now supporting Obama at a 48-45 clip.
Small town voters crossing over from Bush to Obama only account for 2.5% of the sample but even then it's an interesting profile to look at. 100% of them are white. 54% are independents and 38% are Republicans. Sound like an unusual Obama voter? Well 71% of this crew says the economy is its top concern, and that appears to be transcending any cultural differences.
There has also been a major swing among independent Ohioans toward Obama. In 2004 they reported going for Bush at a 51-38 rate. Now they're planning to vote for Obama 48-36, a 25 point shift since the last election.
One thing that seems to be getting lost in the shuffle right now is that young voters really were not that overwhelmingly supportive of John Kerry. The ones under 30 in this survey reported going 48-35 for the Democrat four years ago- they're now supporting Obama 64-28.
There has been movement in Obama's direction relative to John Kerry in every demographic we track.
Full results here.
Tom,
ReplyDeleteWhy do you think Rasmussen has consistenly given McCain overly favorable results in Ohio? All the other pollsters -- Quinnipiac, BigTen, PPP, SUSA -- give Obama leads outside the margin of error in the last two weeks. And big leads in the summer as well. Yet the last Ras poll showed McCain ahead by two and never once Obama ahead in the summer.
I know Scott's a Republican but he doesn't seem to purposely skew his polls.
Is he just weighting his whole sample wrong? Or is it something else?
The thing to note is that 2.5% of the sample is only 25 people. So we have a margin of error of ~20%, and hence we can't draw too many conclusions from them. We can't say with any statistical significance, for example, that those small town voters who are voting for Obama after Bush care more about the economy than the sample as a whole.
ReplyDelete@Tyler, as Tom mentioned in another blog post this week Rasmussen polls are conducted on a single day whereas PPP, Q-Poll etc are conducted over several days....apparently Obama supporters are harder to reach than McCain's supporters.
ReplyDeleteAlso I suspect Rasmussen might be weighting his polls by party ID as well, though I am not a 100% sure of this.
On another note, when will the Virginia poll be released? This is the state, if won, that will make all other obsolete, and I hope the solid margin continues to show itself in your next poll.
ReplyDeleteThe Rasmussen Polls are consistently the best. As a Dem, it predicted 2004 exactly, even though I was not happy. They will be right about Ohio. It will be very close. Obama wins narrowly.
ReplyDeleteRasmussen weights by party id, yes.
ReplyDeleteEvery day done is a good day for Obama. Every day he won more vote because the early voting is huge.
ReplyDeleteAt this moment i think also Colorado is in the bag and Nevada nearly.
Tom,
ReplyDeleteWhen is the Virginia poll coming out? Any hints...