FROM 2012 INTO 2013 POLITICS THREAD

Re: THE 2012 POLITICS THREAD

I'll grant that you may have a point there in terms of tweaks made between elections...I don't know enough about that to speak intelligently on the subject, but I can certainly see that possibility. Where I would find it far-fetched, though, is that each polling firm using its own proprietary methodology would independently come to the same answer, skewed in favor of Obama, across multiple state polls, despite each having a strong incentive to correct for this and prove themselves the most reliable. I.e., if there are mistakes being made that are not capturing Romney's true lead, then everyone is making them across the board despite everyone having their own recipes.

Could there be a herding effect, i.e. everyone comparing their results to the others? Possibly. If there is a systemic bias in favor of Obama, that's the most likely source I can see. But it would require the consensus to be very much off the mark to begin with. And surely we would have seen some of that herd mentality in polling in previous elections.

 

Re: THE 2012 POLITICS THREAD

Couple things to remember:

(1) There may be other aggregators, but there is only one that really matters to right- and left-wingers alike right now, and that's Silver. He's the celebrity prognosticator who works for the New York Times. He's the one everyone follows. No one will give a good goddamned tomorrow night about the prediction of some academic from CalTech or MIT. They'll care about Silver. That's just how it is.

(2) The argument that Silver was right but the polls were wrong will absolutely, positively not wash. There are many, many right-wing voices out there screaming that they know that they polls have a Democratic bias, and that Silver should be factoring it in, as we speak. There will be no excuses for him if Obama loses, as far as the right-wing and independent public is concerned. None.

 

Re: THE 2012 POLITICS THREAD

But don't those firms use the previous election as their baseline? That's one place that I think it all could be thrown off. Young voters and African-Americans came out in never-before-seen numbers. We don't know if they will repeat that effort, and we especially don't know what the effect will be of the GOP's unrelenting voter-suppression efforts aimed at rescinding black people's rights.

 

Re: THE 2012 POLITICS THREAD

Submitted for your approval as to how campaigns in close races get chaotic. Just got a call from a Scott Brown volunteer. I'm a registered Democrat, but since he needs those to win, no surprise. Guy asked for William Gee, name of one of my brothers. I informed him that the only William Gee I knew was a Delaware resident and is out of the country this week anyway. The guy, a nice guy, apologized profusely and hung up. I can't imagine the state of the telephone list the campaign gave this poor devil. I've done phone banking, and I always just hope there aren't too many dead people on the list.
Dick Whitman, I will only point out that while of course it will be awful for Silver if Romney wins, the Gallup Poll actually STOPPED polling Dewey-Truman with two weeks to go because it felt Dewey was a lock, yet over 60 years later, they're still in business and doing well.

 

Re: THE 2012 POLITICS THREAD

Romney - 321 Obama - 217

 

Re: THE 2012 POLITICS THREAD

There is definitely a herding effect as we get to the end of elections. Silver has written about it quite a bit. Anecdotally, a month ago we were seeing anything from 0+5 to R+6. Now everything's within a point or two of each other.

 

Re: THE 2012 POLITICS THREAD

Silver might come out of this looking like a genius, but I do think he's painted himself in a bit of a corner.

If he had 5 states barely leaning towards Obama that Romney needed, it could still lead to him giving Obama an 80% chance to win, since Romney would need to sweep the states.

But, you couldn't criticize him too much if states that he had given Obama a 51% chance to win ended up going to Romney.

The potential problem is that he has Ohio at 86.8%.

And, if a state like WI, PA, or MN -- or all of them -- goes to Romney, he's going to look terrible.

I really don't think that will happen, but that's the risk.

If Gallup and Rasmussen have been right all along, he'll look bad for discounting them in favor of other polls.

 

Re: THE 2012 POLITICS THREAD

Cholesterol levels?

 

Re: THE 2012 POLITICS THREAD

States don't move independently. They don't move perfectly in concert, but it's pretty close.

If the model says there's a 20% each chance of a candidate winning five different states, that's not five different iterations of a 1-in-5 chance. That's five different results being caused by the same 20% situation.

 

Re: THE 2012 POLITICS THREAD

BTW, did Gallup & IBD just give up on their polling because of Sandy?

Other polls resumed. Were they particularly disrupted? Like their offices or employees to the point where it was impossible to go on?

And, won't that make it harder to evaluate polls in the future for folks like Silver?

 

Re: THE 2012 POLITICS THREAD

That's a good point, and you're right, if that were the case, Silver likely wouldn't have it at 80%.

 

Re: THE 2012 POLITICS THREAD

"Look bad" to people who don't understand his model, sure.

He doesn't go into the model and say "Hmm, I don't like Rasmussen, I'll discount them."

Polls are weighted by a pre-decided formula that involves sample size and historical results from that firm.