Not necessarily - other factors, like territory, may come into play. For instance, if you’re polling the People’s Republic of Berkeley, that ratio may be accurate. 
How often do people change party registration for things other than voting in a closed primary? I mean, in the absence of that, would people really switch from Democratic to Republican registration this cycle just because they are voting against Obama? Wouldn’t the most likely scenario be they just keep their registration the same and potentially identify as Republican or Independent instead? Or, if a Tea Party member decides the Republican party isn’t conservative enough, is he really going to take the trouble to re-register as an independent instead of just staying Republican and calling himself independent? Thus, registration isn’t nearly as accurate a barometer of who a population is voting for as party self-identification?
Unless someone has data to the contrary, party self-identification is likely to be far more correlated with voting shares than party registration status. Do any conservatives have any sort of data contradicting this? Otherwise, the argument against the biased polls seems to be wishful thinking. I would be shocked to find that registration status is better correlated, as no major polling company seems to think so, relying on self-identification instead of registration.
Sorry if this is unclear, I’m in a hurry and don’t have time to organize my thoughts better.
Actually, I think most people do change their age from election to election.
It’s a fair cop.
To regginbrow, I honestly have no idea what party (if any) I am registered with. I’ve voted in both party’s primaries in the last decade.
I’m interpretting this as the core of OMG’s argument.
Um, they could both be correct, because fundamentally they are asking different questions. Gallup doesn’t ask what party you are registered under. They ask “In politics, as of today, do you consider yourself a Republican, a Democrat, or an independent?” Different question. Furthermore, the inquiry is pitched near the end of the survey, after people are asked who they are voting for, etc. There’s no reason to believe that Gallup’s question would align with party registration, so re-weighting the poll to make it so is highly dubious.
I don’t have a good take of what the exit polls will show, since I don’t know how they word the party question. And as you noted, early voting and absentee balloting reduces the quality of that exercise anyway.
FWIW 538 upgraded to an 80.8% chance of an Obama victory and a popular vote spread of 2.1 points.
If nothing else Sandy runs down the clock some. (And yes keeps it an Obama possession.)
Yes, they are different questions. But **OMG **was not looking at the *number *of registered Democrats or Republicans or Independents, he was looking at the *changes *in registration, which he posits is a fair measure of the change in party self-ID.
**LinusK **has identified the weak part of OMG’s analysis - the reason Democratic registrations did not go up while Republican ones did is because there was a contested Republican primary, while the Democratic primary was uncontested. Of course there were more people registering as Republican.
Thanks. I should like to see the statistical association between change in party registrations and changes in self-reported party affiliation. I expect a noisy association, even during years where both parties have primaries. Now there could be a relationship, which would imply that there’s a valid method of reweighting the figures. But I seriously doubt whether it is straightforward and furthermore doubt whether it is superior.
If conservatives were serious about this argument they would do such an analysis - the data is out there after all. But I detect a distinct whiff of hackery and wishful thinking. It’s sad really: modern conservatives are becoming more and more detached from reality - this is one of their more plausible arguments after all. They just lack the guts to follow up on it – instead they obsess over Nate’s fashion choices and physique.
Again though: I still think this is a close race, though Obama is favored. In other words, Romney’s chances of victory are between 10 and 40 percent.
Hmm. A tepid endorsement of Obama, but a stinging criticism of Romney.
In order for Obama to win Ohio, Ohio would have to have made a fundamental shift towards Democrats, since Ohio polling shows the President running better in Ohio than he is nationally. That didn’t happen in 2008, in 2008 he ran more poorly in Ohio than he did nationally, by a couple of points.
That’s why Republicans are confident that if Romney wins the popular vote by so much as a point, Ohio’s in the bag.
A 4 point swing in one election is now a fundamental shift? When Ohio in particular has reason to view the auto bailout favourably?
I don’t understand your post at all. You seem to be saying that the Ohio polls must be wrong because we can know a priori that Ohio hasn’t shifted relative to the rest of the nation. But that’s just crazy talk. The evidence shows that it has, though not by very much.
I sincerely hope Obama wins Ohio but I honestly think it will be a miracle if he does. An early Festivus Miracle.
I think adaher is dead on, really, with his assessments.
I’m so scared… somebody… please… hold me!
This is getting a little silly. Ohio almost exactly mimicked the national vote in 2004, but in 2008 the Dem underperformed there by about three points. Was there a “fundamental shift” between 2004 and 2008?
One can easily find similar examples. In 2004 Wisconsin gave the Dem 2.5 points over the popular vote. In 2008, it gave him 7 points over the national vote. Lots of fundamental shifting going around apparently.
There’s no reason to think that a the GOP vote share in a state relative to the popular vote is stable from election to election. Unless that is, your desperately trying to find some argument in favour of Romney winning the upcoming election, and are willing to accept anything that sounds even vaguely plausible without spending the thirty seconds to find out if it holds any water empirically or not.
Look, people. In order for Obama to win Ohio, he is going to have to get more votes here than Mitt Romney. And that is not going to happen. Why, you ask? Well, I will prove it with science.
Obama will not get more votes than Romney in Ohio because that would simply be implausible. Romney getting more votes in Ohio is plausible; Obama getting more votes is not. Q.E.D.
You’re welcome.
If Chewbacca lives on Endor, you MUST acquit!
Well, okay, I’m convinced. Off to invest my life savings on Romney at Intrade.
Why didn’t it help Democrats in 2010? In 2010, Democrats were especially whipped in Ohio. Ohio has shifted from slightly more red than the rest of the country to slightly more blue?
When polls show that something has changed, that change must be acknowledged. If indeed Ohio has shifted more blue, primarily because of greater Democratic enthusiasm than in 2008, that’s a significant change.
However, the polls in Ohio also show something else: that more people are claiming to have already voted than actually have, and the early voters overwhelmingly support Obama. We’ll see if the enthusiasm to falsely claim that they voted translates into actually voting(assuming they are even eligible).
Ohio was more Republican than the country in the last five elections except for 2004, when it was even.
So if Ohio is bluer than the rest of the country, that is in fact significant. Not to mention, Ohio was even more Republican than average in 2010. So yes, going bluer than the country in 2012 would be a significant shift.
I’m going to go out on a limb and make a prediction: Romney wins Ohio 51-48. Reason: The polls in Ohio are skewed by Obama voters falsely claiming to have voted when they haven’t.
adaher, please… my heart just can’t take it anymore.
Could you, just to humor me, please take me back to November 1st, 2004, and tell my why Kerry has got this in the bag and will win it despite (or because of) the polling? I know you voted for Kerry and hated Bush so this should be easy for you.