It’s more volatile, but it was also pretty accurate in the days before 08. And even if it’s volatile, all or virtually all of the polls in WI, NV, OH, and IA have Obama ahead, or at worst, tied. And yes, it’s unlikely that Romney will win the popular vote by 2-3 percent and lose the EC. But not impossible. And it’s probably kind of likely that Romney would win the popular vote by about 1 point and lose the EC- would you really be that surprised?
From Nate, here’sa post from just the last few hours talking about how accurate state polls typically are in the last days before a presidential election.
They’re pretty damn good.
Each individual state poll is more volatile than an individual national poll. However, there are many more state polls than national polls, which more than makes up for this.
Right. When I say (echoing Nate Silver) that the state polls are typically very accurate indicators of what happens leading up to election day, I mean the polls in aggregate- not any particular individual poll.
This isn’t 2008 and, to be honest, most of the polls coming out are laughable. They’ve either got D+7 plus splits and/or Obama leading when he’s losing Independents by 20+ points. Simple logic tells me that neither of those things are apt to happen in 2012, as the electorate is far less favorable for Democrats than it was in 2008. So I’m not too worried about state polling.
I just know that if national trends hold up, Romney will win the presidency.
Ahh, I see. Ignore the polls you don’t like (and complain about voter ID splits)- a recipe for prediction success. If the state trends hold up, Obama will win the presidency, with close to 300 or more EVs.
We’ll know in 10 days. If you’re wrong, I look forward to reading your explanation afterwards for why.
The most recent CNN poll from OH, with Obama up +4, shows D+3 voter ID (just like you predict nationally). That’s in line with most of the recent OH polls.
It really does seem to me like you’re ignoring good data that conflicts with your narrative.
It worries me that the Romney family has a financial interest in some of the voting machines in Ohio.
http://www.opednews.com/articles/Rigged-Elections-for-Romne-by-Michael-Collins-121022-13.html
It worries me that voting machines are easy to hack into and often leave no paper trail.
It worries me that employers are turning the screws on their employees in favor of the GOP in this election. Since that sort of intimidation hasn’t happened in this country in generations, who knows what the effect might be?
http://finance.yahoo.com/news/ceo-workers-youll-likely-fired-131640914.html
It worries me that right wing religious leaders are threatening damnation for Obama voters.
http://www.addictinginfo.org/2012/10/26/another-catholic-bishop-threatens-damnation-if-congregation-votes-for-obama/
It worries me that hundreds of millions of dollars scurrilous ads are being run:
http://www.nytimes.com/2012/10/24/us/politics/strident-anti-obama-messages-flood-key-states.html?_r=1&hp
It worries me that only some of the Republican voter registration fraud has been caught and stopped.
It worries me that hatred might be a greater motivation for voting against Obama, than love of something constructive is in voting with the president.
Let’s take Ohio, for example.
In 2008 exit polls show Obama won Ohio by a little under 5% while winning Independents by about 8%. Now, if you look at that graph, not only is Obama losing Independents in Ohio this time by about 10.4%-- a near 19% difference-- but if most of those polls are to believed, Obama is either going to nearly triple his turnout in 2008 or Republican turnout is going to plummet to record low levels even though the Ohio electorate in 2010 was R+1. Neither of those things is likely to happen, especially considering that Ohio is generally a bit more Republican than the nation as a whole.
And since I’m not looking at any of them, it’s not cherry picking (as if I were I’d simply look at the good ones).
In the CNN poll I linked to, Obama’s winning independents.
But it’s really weird how you focus on Voter ID, despite many posters here and even the pollsters themselves, like Gallup, insisting that it’s a very fluid measure and shouldn’t be focused upon. The only way to predict how people will vote is to poll and ask them. That’s what’s going on, and in the states, the polls are showing Obama with a lead. There’s lots of other demographics that could be looked at- like how Gallup thinks minorities will turn out- that look very odd as well. But that just may be how their respondents are telling them.
Nate Silver was very, very close in 08, he was very close in 2010, and I think it’s very likely he’s going to be close this election. You’ve completely failed to show any weakness in Nate’s methodology. The scenario Dan and you throw out that results in a Romney +2 popular vote result is very possible. But tweak your assumptions just a tiny bit- +1% more Democrats, and -1% fewer Republicans, and the election is dead even- and if you’re off by just a bit more, then Obama’s ahead. I think your and Dan’s scenario is at the optimistic end (for Romney) of what’s likely to happen. I think Nate’s prediction, of about +1 for Obama, is about in the middle of what’s likely to happen, with a smaller likelihood of a larger Obama victory. And in the states, Obama has a greater advantage according to polling.
Now who’s cherry picking polls? The link I gave you includes the CNN poll in the average.
Party identification isn’t like underwear that you (hopefully) change every day. Though that’s something many pollsters would have you believe, with large variations from one day to the next (it’s unreasonable to think that people switch between Democrat and Republican daily). Furthermore, this hinges upon the assumption that the likelihood of any specific person of answering the phone is about equal. If, for examples, Democrats are more likely to answer a survey (I’ve no way to prove this as true; just throwing it out there), then you will get an inaccurate picture of any given race as while the internals might be correct, the topline will not be.
Because he doesn’t have one. His weighting system is completely arbitrary. And I do believe he was given internal polling data from the Obama administration in 2008.
D+3 isn’t optimistic for Romney. Optimistic for Romney would be even or R+1. D+1 would be an assured Romney victory. As it is, Obama is not likely to get D+5 without a depressed Independent vote.
“It doesn’t matter what someone predicted years ago. Being wrong in the past does not preclude you from being right today, anymore than being right in the past doesn’t preclude you from being wrong today.”
It matters *why and how *they were wrong, if they repeat the same process today. I certainly don’t see any objective reason to think Dan is more accurate than Nate, or CNN in Ohio. Why was he so wrong 4 years ago? I would propose it is because he was partisan, and **making an argument ** (to quote from your OP) for why McCain would win, rather than following the evidence where it led him. How do you rate the liklihood that Unskewed polls will call it accurately?
(After I typo my header, can I fix it within the editing period?)
I have a question…is the increasingly irrational argument from OMG leading up to some kind of Andy Kaufman-esque grand prank? Because I just want to know if there’s going to be a payoff before I invest any more time or effort into actively scorning it.
Perhaps, though you really don’t know. That’s why we look at lots and lots of polls together. Every election the losing side, in the waning days, cherry-picks trends and details in the polls that they say don’t match reality. Every election. Because every election is different, and every electorate is different, so every 4 years there will be something that looks “strange” compared to the last one.
I think it’s likely that your side is doing this once again- just like they did in 2008, and just like my side did in 2004. The polls say X. It’s more likely that X is going to happen, or something close to it, then that the polls are wrong by a significant amount, in aggregate.
He doesn’t give out the details of his proprietary system, but we know that it’s been good in 08 and '10. Whether it’s arbitrary or not (and whether he gets data from Obama or not- and how is that a bad thing?), it’s been quite close to dead-on. Historically it’s not a good bet to bet against Nate.
Well we really don’t know for sure, we can just estimate. I think your and Dan’s estimates are at the optimistic limit for your side. I think it’s very possible that relative to the whole population, there are just not as many Republicans as you think they are- whether it’s because of demographic changes (more minorities and young people then in 08) or because the Republican party has lost popularity.
We’ll see. I’m really looking forward to it.
Is this the sort of thing you’re worried about?
http://www.addictinginfo.org/2012/10/27/gop-rigging-elections-for-romney/
Been hearing rumbling about this, but don’t know if there’s anything to it.
Why bother, at this point? We’ll find out in nine and a half days. In the meantime, I doubt that anyone is going to change anyone’s minds around here about who is more likely to win, not that it would matter if you did.
If it makes you feel better to believe Romney’s really more likely to win next week, then who am I to piss on your belief?
Some beliefs just need pissing on.
you know, that’s a pretty good motto for my life…
As far as i am concerned Nate Silver is right until he is wrong, and we won’t know that for a few more days.
My understanding is that Mr Silver, that effeminate statistical version of Percy Dovetonsils, has incorporated the effect of voter suppression into his formula, and still predicts a slight advantage to Democrats.