Final Stretch Election ThreadFollow

At this point in the campaign, it'd be nice to hear either side hold a rally or whatever and say "You know what, we don't need any more of your cash. We've already raised a billion dollars this cycle. Why don't you give that right to the Red Cross [or some other reputable charity]?"

Then after they lost the election, we'd remember the fondly probably until Thanksgiving.

Omega,
Were it that easy. I am in an undisclosed location where purchasing things like Premium is not inconvenient, but actually impossible. Moreover, I am only able to view this in HTML mode so I have no idea what your avatar is, thus will be unable to pick one out for you come Tuesday night. As it is, I am fortunate that the internet is not provided for me by smoke signals and semaphore flags.

However, a sig is something I can do. Let me know.

Totem

It's like you're reading my mind. It's not just oversampling, but also overweighting of democrats (depending on poll methodology). There's an 11 point shift in terms of party affiliation from Dem to GOP between 2008 and 2012, but many polls are weighting results based on an assumed turnout similar to 2008. While not all polls do this, even among those which don't some are reporting poll participation rates by Dems much higher than would be expected based on identity polling. Which either means that Dems are being over represented in one set or under represented in another (or some combination of the two). Given the trend based on 2010 figures, the former is more likely than the latter.

How much this affects the overall polling picture (like RCP) is hard to say, but I really wouldn't assume those numbers are accurate. Same thing happened in 2008 (only in reverse), where polling indicated a much closer race between McCain and Obama than ended out happening. Obama's slim leads in polling in key battleground states could easily really be large losses when it comes to election day. We'll find out in 5 days, I suppose.

Well, obviously.

None of the Ohio polls have a 5pt MoE. It's closer around 3-3.5pts. But anyway, that's why you look at them in aggregate. When you have ten polls all showing an Obama lead and averaging, say, O+2.5 the chance of them all being wrong on the far edge of the MoE and the real answer being R+2 is pretty minimal.

Edited, Nov 1st 2012 5:09pm by Jophiel

Yeah, pretty much. It really depends on the poll though. Some are pretty consistent, so you can track changes and say reliably that so-and-so is doing better/worse than a month ago. Others swing widely, so it's not just the margin of error you have to look at but longer term trends in said poll. But both kinds are only semi-effective at actually telling us how an election might turn out. And in this election, the methodologies are presenting us with a really big factor that usually isn't so huge. Obama energized a *huge* number of people in 2008 to turn out and vote Dem. Normally, the participation rates don't change much between elections, but it did that year. You really have to consider it an outlier in terms of Dem turnout. But polls don't change their methodology to account for this (because it's important for them to continue using the same formula every time, for obvious reasons).

You can see this when you look at RCP average polling. You'll see about half that show the candidates within a couple point spread (some with Romney up, some with Obama up, but all close), and another half that have Obama up by 5-8%. When averaged, those give a 2-3 point lead for Obama. But garbage in, garbage out, right? You're generating an average where a significant portion of the data being used is quite obviously not going to be accurate. But those are the numbers, and those are the methods used, so they continue to use them.

Can't measure the exact effect of this, but you could almost safely say that any poll that uses this methodology (at any point in their process) is likely to be off by 5+ points. You could even stretch that (if you wanted to really crunch numbers) and speculate that they could be off by as much as 10-12 points. This could result in anything from a very tight win for Romney, to a landslide for him. In any case, it's almost certain that a whole lot of people are going to be surprised by the actual election results and wonder why they were so different from the polling data they were counting on being true.

Do you even look at the RCP site or do you just mindlessly parrot whatever someone else tells you? RCP has a single poll since Oct 10th in Ohio with Romney up (a Rasmussen poll from a few days ago). Every other poll out of the nine in the main calculation right now has Obama leading. Obama's largest lead in an Ohio poll from the main calculation is +5.

Seriously, there's no way you could honestly be looking at the numbers you try to speak so authoritatively on and say the things you say.

Unless you're responding to a tweet about Ohio polling by talking about polling in... umm... Montana or something. I dunno.


You would be best to ignore Gbaji on this topic. Last time he was swearing that no one was using Likely Voter screens when only a single pollster out of eleven (Gallup) wasn't using one. Gbaji literally has no idea what he's talking about and would rather just bluster statements than take half a second to look at a site and verify what he's trying to tell you.

Edited, Nov 1st 2012 5:24pm by Jophiel

In Romney's defense, that's not exactly difficult to do. We're a hot blooded breed.

If you're going to do a Jar-Jar impression, wouldn't my Star Wars thread would be a more suitable venue?

You can call cell phones for polling but it has to be a live person calling. You're not allowed to robo-poll cell phones (which means polling firms like Rasmussen miss them).

They still do, my roommate interned at a polling station this past spring.

But it is true that it has become much more difficult to get a representative sample, because more and more people are disconnecting their home phones, and it's much harder to get access to cell numbers. And you don't want to include both, because it would be a serious issue if you double listed 1/3 of the people in your set.

So, people are telling Republicans "I totally pressed the Romney button but it says I voted for Obama instead!"

I wonder how many of them had their fingers crossed when saying that.

It could have been a misconfiguration, definitely. And it could be displaying Obama at the end review, but still recording the vote for Romney, because someone copied and pasted the code wrong.

For all I know, it's true. Seeing as how I've been against fully digital systems for years

Why? Aside from random "I'm a control freak and I understand how paper ballots work, but not really digital ones therefore NO ONE CAN!" issues. Why in the world would you think it was harder to game paper ballots. Every digital voting issue is alarming ONLY because of it's novelty. There are literally thousands of ways to abuse paper ballots. If you're concerned about fraud, the absolute only viable solution is non anonymous balloting. Otherwise, every system is gamable, and gamable systems will be gamed. How frequently this affects outcomes is another issue entirely. It's fairly likely the answer is almost not at all.