Kilmer17s non confrontational election prediction-

[Tulane Skins Fan]

This is a good article, IMO:

http://www.redstate.com/2012/10/31/on-polling-models-skewed-unskewed/

Edit:

Here's another non-scientific take on why they're wrong. Long-story short...intensity.

http://www.nationalreview.com/articles/332611/polls-ate-presidential-contest-mona-charen

Probably the first article I've read on redstate.com

Its an interesting debate, but I didn't find him very convincing. Basically, he still says that he "feels" things are going better for republicans these days.

He could be right though. We'll all find out tonight and then we can see how the polls were right or wrong though. I think Obama's gonna win, but I wouldn't give myself any 90% confidence like Silver does. More like 60% confidence.

[Wrong Direction]

I think this election will have a profound affect on how polls are used and abused going forward.

I think the methods are really tough to work through. More early voting in particular is a new factor. It's hard to account for all of that stuff.

---------- Post added November-6th-2012 at 02:01 PM ----------

I think Obama's gonna win, but I wouldn't give myself any 90% confidence like Silver does. More like 60% confidence.

Silver's 90% prediction is a math prediction, not likely his personal beliefs. If the polls are wrong/turnoff exceeds in 1/2 states, Romney becomes the favorite. It's just too close to know.

[MattFancy]

Probably the first article I've read on redstate.com

Its an interesting debate, but I didn't find him very convincing. Basically, he still says that he "feels" things are going better for republicans these days.

He could be right though. We'll all find out tonight and then we can see how the polls were right or wrong though. I think Obama's gonna win, but I wouldn't give myself any 90% confidence like Silver does. More like 60% confidence.

Silver's getting his numbers from polls. So we all know they could be off, but even still, I'd put Obama's chances at about 70% right now. I think he is definitely the favorite. I still believe he ends up somewhere near 280. I don't quite think its as close as some believe, maybe the popular vote will be very close, but I think Obama is sitting pretty good right now with the EC numbers. I could very well be wrong, but that's my feeling.

[Heisenberg]

Based on the polls I have O winning with 303 EV's.

I'm basing that off most of the polls though, so I have no real idea what to expect since it seems like this year is different than most.

[MattFancy]

Based on the polls I have O winning with 303 EV's.

I'm basing that off most of the polls though, so I have no real idea what to expect since it seems like this year is different than most.

Well according to CNN's Poll of Polls, it has Obama leading before today in Ohio, NH, Virginia, and Wisconsin. Now that was as of 11/1 so it could be different now, but if that's the case, that would give him around 294 or so.

[drtdrums]

Presupposition 1: Higher enthusiasm for Romney than McCain

Presupposition 2: Lower enthusiasm for Obama than 2008

Presupposition 3: Any poll at more than D+3 is worthless (D+7 and higher have been the trend among those showing Obama leading by any margin).

Presupposition 4: Romney is going to hold his large lead with independents

If all of these are correct, Romney wins.

If 3 of these are correct, Romney wins.

The early voting numbers for Ohio are EXTREMELY positive for Romney. If what has been reported is true, I see no possibility of victory for Obama there.

Reports are also that there is lower turnout in Madison, WI. Obama NEEDS a huge turnout there, as WI is entirely dependent on Madison's blue flood. If the students didn't all just sleep in, WI looks like a Romney win.

PA...I dunno. And no one else does, either.

NH, the internals look extremely encouraging for Romney.

It appears FL, NC, and VA are all going to Romney. Virginia polls showing Obama tied or slightly ahead all suffered from D+7 syndrome.

Colorado is almost certainly Romney.

Iowa was leaning Romney at last polls.

....

Basically, the only way Obama wins is with 2008 turnout or just *slightly* less. I just can't see that happening.

Today's info is looking more and more like 300 EV's for Romney, at least early. If I'm wrong, I'm wrong because Dems do REALLY well this evening, and that would buck the usual Republican evening trend.

[Duckus]

As for my prediction, I think the national polls are slightly - and wrongly - skewed toward Obama because of a small Superstorm Sandy-Christie bump that ended on Sat/Sunday. I think Romney wins by 1 to 1.5 points in the pop vote. If Romney outperforms, he could win by 2-3 pts. If Obama outperforms, I think he only wins by <1% of national vote.

Of the swing states, I think Obama wins Michigan, Nevada, Wisconsin, Minnesota, New Hampshire and Pennsylvania. I think Romney wins Colorado, Florida, Iowa and Ohio.

In Ohio, the early vote "lead" by Obama is much less than 2008 and generally means Romney needs about 52% of the voters today in order to win. I think many R's are pretty confident he'll get that 52%, with many polls predicting Romney wins 54-55% of election day voters (remember, a lot of Ohioans vote before election day). Ohio has outperformed for R's versus the rest of the nation for decades. I'm not sure why it would be different today.

This all gives Romney a narrow EC win and a narrow popular vote win.

If I'm wrong, I blame twitter and facebook...and I'm actually serious. If the polling turnout models are right, I think that means they reflect a new and sustainable participation rate among Democrats. If you follow twitter or facebook, you see more about voting than ever before. I suspect this drives turnout for both parties above historical levels, but there's more room for growth in Democrat constituencies. This could mean Obama wins Virginia in particular, in addition to Ohio.

If I'm right, it's because young, hispanic and traditionally swing voters have all really abondoned Obama. They won't necessarily go for Romney, but they will have stayed home. Conversely, evangelicals and independents would drive a Romney turnout above and beyond what state pollsters in particular are finding. Wisconsin and Pennsylvania could both go to Romney with a relatively small "over" performance.

Nice post.

If Romney does win I agree and think he takes Colorado, Florida, Iowa, and Ohio like you predicted. Having said that, I don't think it will happen, but it is possible if there is a catastrophic breakdown of polling.

I don't think if Obama wins social media is not THE reason. Yes, Obama's campaign is amazing at social media, but the real Obama advantage is 1) data (I can't stress this one enough) and 2) having a campaign infrastructure on the ground for 8 straight years in many states.

[Heisenberg]

drtdrums, I'm sure what you are saying will happen is possible. However, I'm not sure what evidence there is that points to any of that being likely.

I don't think Obama has ever trailed in Ohio in the aggregate polls.

[Predicto]

I read a lot from people that I think are smart (left and right) and I sense uneasiness on both sides. I think Silver is flawed, but he does reflect the reality of the polls. As a R, I can't ignore that reality and I can't like that the reality from polls points to a small Obama win in a lot of places. On the other hand, if you've read the case against many of the polls models, it raises a strong possibility that the polls really aren't that good.

Either way, this is close. This election could fall within the margin of error of most polls and still go either way. That's why we have elections. Turnout will be the decider. That is the #1 source of confidence for Rs these days.

Regardless of who wins, I sure hope it is clear tonight or early tomorrow. No recounts, please.

I agree with all of this.

[Kilmer17]

drtdrums, I'm sure what you are saying will happen is possible. However, I'm not sure what evidence there is that points to any of that being likely.

I don't think Obama has ever trailed in Ohio in the aggregate polls.

While true, the determining factor today will be whether or not the weighting used in the polls is correct.

If Dems are oversampled by 7 and Obama has a 3 point lead, that would mean Romney is probably winning. Furthermore, 5 out of the last 7 were in the margin of error.

[Duckus]

Presupposition 3: Any poll at more than D+3 is worthless (D+7 and higher have been the trend among those showing Obama leading by any margin).

Presupposition 4: Romney is going to hold his large lead with independents

In my opinion, people don't seem to understand the logical connection between 3 and 4.

Polls are finding more Democrats when conducting polls. Why? As it has been reported, more traditional Republicans are calling themselves Independents rather than Democrats in recent years. You see that trend even here on ES, with many strong conservative posters consistently calling themselves Independent.

This has two effects: 1) It makes Democrat margins looks crazy high in polling 2) It makes Romney look like he is kicking ass with Independents.

But that is a theory, and we will see who is right in a few hours.

[Kilmer17]

In my opinion, people don't seem to understand the logical connection between 3 and 4.

Polls are finding more Democrats when conducting polls. Why? As it has been reported, more traditional Republicans are calling themselves Independents rather than Democrats in recent years. You see that trend even here on ES, with many strong conservative posters consistently calling themselves Independent.

This has two effects: 1) It makes Democrat margins looks crazy high in polling 2) It makes Romney look like he is kicking ass with Independents.

But that is a theory, and we will see who is right in a few hours.

That's not how weighting of polls works.

[Duckus]

That's not how weighting of polls works.

I don't want to go down the rabbit hole today of poll weighing. It is just not done in a way that most people understand.

We will agree to disagree.

[Heisenberg]

While true, the determining factor today will be whether or not the weighting used in the polls is correct.

If Dems are oversampled by 7 and Obama has a 3 point lead, that would mean Romney is probably winning. Furthermore, 5 out of the last 7 were in the margin of error.

The aggregate polls would usually help to reduce any oversampling in the polls - in the past anyway.

Like I said, I guess it could be different this year. However, I don't recall any situation in the past where the pre-election polling was so off at this stage to produce the results that some seem to be anticipating.

Then again, it only takes a state or two to flip unexpectedly so who knows.

[Bliz]

In Ohio, the early vote "lead" by Obama is much less than 2008 and generally means Romney needs about 52% of the voters today in order to win. I think many R's are pretty confident he'll get that 52%, with many polls predicting Romney wins 54-55% of election day voters (remember, a lot of Ohioans vote before election day). Ohio has outperformed for R's versus the rest of the nation for decades. I'm not sure why it would be different today.

The answer is the auto bailout. All indicators are that it has played extremely well for Obama in the rust belt. That Jeep ad brought it right back up to the forefront of the conversation just before the election, and it was big news when the company said the ad wasn't true. If Ohio is different today as far as outperforming for Rs vs the rest of the nation, I think that's the reason. Or at least a big part of the reason.

The early voting thing is interesting. Someone (can't remember who) on POTUS radio on XM this morning was saying that what he saw was indeed a much smaller lead for Obama in early voting in Ohio, but largely owing to increased Republican participation in early voting, not a decrease in Dem participation. If that's true (and it would make sense - McCain wasn't nearly so focused on the early vote operation in 2008 as Obama was, and Republicans learned their lesson after last election) then the more narrow lead might be a wash. I haven't seen anything before about polls of election day voters.

[Kilmer17]

I don't want to go down the rabbit hole today of poll weighing. It is just not done in a way that most people understand.

We will agree to disagree.

I realize that. But you're making the claim that polls are finding more Dems than Rs, and then using THAT as the sample to weigh. That's not what they do. They already assume the bias in Ind who are really Rs. What they do next though is the key piece. They take their total sample, and THEN apply a mathmatical formula to accurately affect what the pollsters thinks will be the electorate split.

If Nate Silver is right, and this election is closer to 08 than 04 in terms of the number of Obama voters showing up compared to Romney voters, then Obama will win handily.

If his weighting model is wrong, and the election is closer to 2010 in terms of the turnout at the polls, then Romney wins in a landslide.

BTW, this is one of the reasons the leftwing used to ***** and complain about Rasmussen all the time. His models always took a sunny outlook for the GOP early on, and then, as his dailies get fine tuned as the election gets closer, he changes to reflect what he sees in CURRENT polling. Silver has remained static since the beginning.

[Kilmer17]

The aggregate polls would usually help to reduce any oversampling in the polls - in the past anyway.

Like I said, I guess it could be different this year. However, I don't recall any situation in the past where the pre-election polling was so off at this stage to produce the results that some seem to be anticipating.

Then again, it only takes a state or two to flip unexpectedly so who knows.

Take a look at the 08 polls. Most had Obama winning Ohio by 2-3 (RCP av). But they were using 04 modelling to set those weights. It turned out he won by 8. That's a simple expression of why people are questioning the way some polls are being weighted.

[Heisenberg]

I realize that. But you're making the claim that polls are finding more Dems than Rs, and then using THAT as the sample to weigh. That's not what they do. They already assume the bias in Ind who are really Rs. What they do next though is the key piece. They take their total sample, and THEN apply a mathmatical formula to accurately affect what the pollsters thinks will be the electorate split.

If Nate Silver is right, and this election is closer to 08 than 04 in terms of the number of Obama voters showing up compared to Romney voters, then Obama will win handily.

If his weighting model is wrong, and the election is closer to 2010 in terms of the turnout at the polls, then Romney wins in a landslide.

BTW, this is one of the reasons the leftwing used to ***** and complain about Rasmussen all the time. His models always took a sunny outlook for the GOP early on, and then, as his dailies get fine tuned as the election gets closer, he changes to reflect what he sees in CURRENT polling. Silver has remained static since the beginning.

Static?

He had the race at <60/40 just a few weeks ago.

[Heisenberg]

Pretty sure Ras comes closer to what the other polls are saying as the election approaches to try to keep any sort of credibility.

But whatever, another area where we can agree to disagree.

[Kilmer17]

Static?

He had the race at <60/40 just a few weeks ago.

He's been static with his modelling, not in his evaluation within the models. He's maintained a large D advantage and hasnt changed it.

[Tulane Skins Fan]

I realize that. But you're making the claim that polls are finding more Dems than Rs, and then using THAT as the sample to weigh. That's not what they do. They already assume the bias in Ind who are really Rs. What they do next though is the key piece. They take their total sample, and THEN apply a mathmatical formula to accurately affect what the pollsters thinks will be the electorate split.

If Nate Silver is right, and this election is closer to 08 than 04 in terms of the number of Obama voters showing up compared to Romney voters, then Obama will win handily.

If his weighting model is wrong, and the election is closer to 2010 in terms of the turnout at the polls, then Romney wins in a landslide.

BTW, this is one of the reasons the leftwing used to ***** and complain about Rasmussen all the time. His models always took a sunny outlook for the GOP early on, and then, as his dailies get fine tuned as the election gets closer, he changes to reflect what he sees in CURRENT polling. Silver has remained static since the beginning.

I don't mean to really open up a can of worms, but I think (emphasis on "think") that you are misunderstanding what Nate Silver does. I don't believe he weighs the polls based upon likely voter models or d's v. r's like you are indicated. I think he weighs the polls based upon past performance accuracy, size of polling data, and other factors. But I don't believe he changes any of the numbers that come to him.

For instance, I don't think he gets the Rasmussen poll which says its 48-48 and then reweighs that poll and adjusts the numbers to make it 50-46 Obama. I think he takes the numbers of Rasmussen and weighs it relative to other polling. I think that's a major difference between what you said and what he does.

Of course, Silver doesn't tell us his formula so I guess none of us really know. But I haven't ever read him saying that he's re-adjusting weighted polls.

[Duckus]

I realize that. But you're making the claim that polls are finding more Dems than Rs, and then using THAT as the sample to weigh.

No, that is not what I am saying at all.

As I have said before, 99% of pollster don't weigh on Party ID at all. They weigh on demographic factors and allow Party ID to be fluid. My point is that starting a few years ago less and less people call themselves Republicans and instead started calling themselves Independents. This makes Party ID seems biased to Democrats in polling results and also gives a false lean of Independents to the right.

Both good sources of what I am trying to say:

http://www.huffingtonpost.com/nick-gourevitch/romney-lead-with-independents_b_2058290.html

http://talkingpointsmemo.com/archives/2012/10/can_romney_win_indies_lose.php

[Bliz]

He doesn't publish the formula for the model. How would you know if he has been static or not?

[Kilmer17]

Not a can of worms at all.

Regardless of whether he changes or not, if he is using polls that already overweigh one side, it will produce false results.

Silver predicted 49 out of 50, but while the polls were right about who won, they were more often than not wrong about the margins.

[Prosperity]

Nate Silver is not a polling agency, obviously he doesn't weigh the samples

He compiles polls and weighs the value of the polls

I'll go further and say that the (shady) practice of changing methodology midstream is taken account because the polls are taken into account. And since Rasmussen changes their methods, and since Rasmussen is one of the polls considered by Silver, then obviously the change in Rasmussen's methods effects Silver's results.