2012 Presidential Campaign

[Easy Mac]

Originally Posted by Buccaneer

I still don't see it. It's like being down 3 games to 1 and needing to sweep to win.

Having the Executive and Senate of the same party is a very bad idea and I wish it would be different.

Just for fun (while I'm waiting for the database to come back up), here are my one-word answers for each of the administrations in my lifetime:
Obama: Amateurs
Bush2: Arrogant
Clinton: Self-centered
Bush1: Detached
Reagan: Hype
Carter: Atrocious
Ford: Transitional
Nixon: Corrupt
LBJ: Tumultuous
JFK: Facade
Ike: Traditional

I didn't know it was possible, but Bucc is both older and younger than I imagined.

[Crapshoot]

My guess on the Senate / imp races:

MA: Warren rides Obama coattails, even if she drags behind him significantly.
VA: Kaine over Allen, even as Romney wins the state by a whisker.
AZ: Carmona gives him a scare, but Flake holds on to win.
MO: Akin does better than expected, but still loses.
IN: Joe Donnelley just barely pulls it out, but this is the one I have the least confidence in.
MT: Tester wins barely.
ND: Berg pulls it out; Herkamp's Obama headwind is too much.

[JonInMiddleGA]

Originally Posted by panerd

I would disagree that a big chunk of the GOP believes in smaller government. They just oppose the left's big government ideas. They fully support the Patriot Act, endless war in the Middle East, health care expansion when it is their idea, actually to piggyback on that pretty much any expansion of the government when it is their idea.

I would not disagree with this general take. For me it's considerably more about "what" than "how much". I would stop short of saying that's a majority view of GOP voters but it's a sizable component.

[Solecismic]

Originally Posted by kcchief19

I don't know that he's leaning toward Democrats being seriously underpolled. Maybe +1, but not much more than that.

Silver initially defended Rasmussen polls in 2010 before turning on them when they underperformed in the midterms.

I don't think Silver's mistrust of Rasmussen polls are political but rather methodical. Rasmussen's polls took a big turn toward Republicans in 2010, and Scott Rasmussen started getting cozy with some key Republicans. Rasmussen is a businessman, not a pollster. I think the perceptions by other pollsters is that Rasmussen uses poor methodology.

All polls biased in one way or another. Whether through sampling data, quotas or screening most polls we see have some "hand of god" influencing them. Typically this is done to fit a profile -- Gallup thinks they know what a likely voter looks like, and no one is going to change their mind, no matter how bad they get the results most of the time.

If you've read any background on the creation of ESPN and the Rasmussens, you'd probably feel very differently about their polls. Scott Rasmussen is a business man, and I have no doubt he'll tilt his poll where the money is.

It sounds like Rasmussen deserves some criticism, and I can understand why his results would be adjusted. In my case, I'm adjusting the result itself, in Silver's, he's giving it far less weight.

You can look at Silver's analysis before 2010, and he praises Rasmussen. Since then, I was picking up skew without knowing why, but it might be more.

My complaint is that Silver isn't challenging his own employer. The CBS/NYT polls seem so skewed I'm not even using them. I looked into them, and found a D adjustment that can't be defended (+9 last month when I looked). Silver called the CBS/NYT one of the worst in 2010, yet, now that he's paid by them, he's giving it his strongest weight.

This doesn't invalidate Silver's results or methodology. But it does call into question a bias.

My question to you and others is, given you have a point about Rasmussen, why are similar criticisms of polls that favor the Democrats defended so vigorously? Why can't this work both ways? Is your side always right, no matter what?

[larrymcg421]

So I keep hearing about these evil D+9 polls that are making the race closer than it appears. Here is the spread for each of the major polls aside from Gallup and Rasmussen which I could not find (and I don't think anyone is arguing they are D+9):

PPP (D): D+2 (Romney +1)
CBS/NYT: D+5 (Obama +1)
ABC/WP: D+5 (Romney +1)
Pew: D+1 (Tie)
NPR: D+6 (Romney +1)
IBD/Tipp: D+7 (Obama +1)
Reuters/Ipsos: D+6 (Obama +1)
YouGov: D+1 (Obama +1)

[RPI-Fan]

Originally Posted by Solecismic

It sounds like Rasmussen deserves some criticism, and I can understand why his results would be adjusted. In my case, I'm adjusting the result itself, in Silver's, he's giving it far less weight.

You can look at Silver's analysis before 2010, and he praises Rasmussen. Since then, I was picking up skew without knowing why, but it might be more.

My complaint is that Silver isn't challenging his own employer. The CBS/NYT polls seem so skewed I'm not even using them. I looked into them, and found a D adjustment that can't be defended (+9 last month when I looked). Silver called the CBS/NYT one of the worst in 2010, yet, now that he's paid by them, he's giving it his strongest weight.

This doesn't invalidate Silver's results or methodology. But it does call into question a bias.

My question to you and others is, given you have a point about Rasmussen, why are similar criticisms of polls that favor the Democrats defended so vigorously? Why can't this work both ways? Is your side always right, no matter what?

The criticisms of the the polls favoring the Democrats are not similar. The fundamental criticisms of Rasmussen are, most simply put: (1) he does not call cell phones; and (2) he weights his results based on how he thinks the party identification of the electorate will look on election day.

I assume the criticisms of the Democrat-leaning polls are that they improperly weight the basic demographic makeup of the electorate. While this may be a valid attack, it is far, far different from what Rasmussen does. Weighting by demographics is well-accepted as a necessary and sound technique, though there is certainly potential for doing it wrong (which is where I assume the criticisms of the Dem-leaning polls are directed). Weighting for party identification is thought of as a completely unsound polling methodology. Perhaps Rasmussen will be vindicated in this election, and weighting by party ID will become the norm. But it is unsound science based on the state of the art at this time. So it is much different to criticize him for doing something so widely panned than it is to criticize other pollsters for using accepted techniques but disagreeing with the way they do them.

[Crapshoot]

Originally Posted by RPI-Fan

The criticisms of the the polls favoring the Democrats are not similar. The fundamental criticisms of Rasmussen are, most simply put: (1) he does not call cell phones; and (2) he weights his results based on how he thinks the party identification of the electorate will look on election day.

I assume the criticisms of the Democrat-leaning polls are that they improperly weight the basic demographic makeup of the electorate. While this may be a valid attack, it is far, far different from what Rasmussen does. Weighting by demographics is well-accepted as a necessary and sound technique, though there is certainly potential for doing it wrong (which is where I assume the criticisms of the Dem-leaning polls are directed). Weighting for party identification is thought of as a completely unsound polling methodology. Perhaps Rasmussen will be vindicated in this election, and weighting by party ID will become the norm. But it is unsound science based on the state of the art at this time. So it is much different to criticize him for doing something so widely panned than it is to criticize other pollsters for using techniques but disagreeing with the way they do them.

"".

[JonInMiddleGA]

And at the end of all the polls, it basically boils down to who had the highest percentage of accurate responses that turn into actual voters.

[RPI-Fan]

Originally Posted by JonInMiddleGA

And at the end of all the polls, it basically boils down to who had the highest percentage of accurate responses that turn into actual voters.

This is certainly true. And if you assume that errors on this score are distributed evenly across the spectrum of polls out there (they may well not be -- perhaps this is finally the 21st century election where Gallup's 20th-century likely voter screen hits the mark), then Obama takes Ohio, Wisconsin, and Nevada rather easily and wins a second term.

[sterlingice]

The good news about all this poll noise is that it's an open book test. We'll see all the predictions the day before and day of and we'll then see the results so we can see who did what

SI

[albionmoonlight]

Think of the Children:

Tired of Bronco Bamma and Mitt Romney - YouTube

[Passacaglia]

Originally Posted by RPI-Fan

This is certainly true. And if you assume that errors on this score are distributed evenly across the spectrum of polls out there (they may well not be -- perhaps this is finally the 21st century election where Gallup's 20th-century likely voter screen hits the mark), then Obama takes Ohio, Wisconsin, and Nevada rather easily and wins a second term.

But if I assume the errors on this score are distributed the way I prefer them to be -- huzzah!!

[M GO BLUE!!!]

Ex-FEMA director Michael Brown criticizes Obama for reacting too quickly to storm | The Lookout - Yahoo! News

Former FEMA director Michael Brown, who was heavily criticized for the agency's failure to prepare for Hurricane Katrina, has criticized President Obama for responding to Hurricane Sandy too early.

[larrymcg421]

Forgot to include this on my earlier post about partisan split in polls. The evil PPP's latest poll has only a D+2 spread and a Romney +1 result.

[RendeR]

Originally Posted by M GO BLUE!!!

Ex-FEMA director Michael Brown criticizes Obama for reacting too quickly to storm | The Lookout - Yahoo! News

Former FEMA director Michael Brown, who was heavily criticized for the agency's failure to prepare for Hurricane Katrina, has criticized President Obama for responding to Hurricane Sandy too early.

I find it hard to believe anyone gave him time at all. He offers nothing to anyone anywhere. he is the epitome of failure at the highest level.

[molson]

Originally Posted by M GO BLUE!!!

Ex-FEMA director Michael Brown criticizes Obama for reacting too quickly to storm | The Lookout - Yahoo! News

Former FEMA director Michael Brown, who was heavily criticized for the agency's failure to prepare for Hurricane Katrina, has criticized President Obama for responding to Hurricane Sandy too early.

I thought that was an Onion article until I clicked on the link.

[JPhillips]

Originally Posted by Solecismic

It sounds like Rasmussen deserves some criticism, and I can understand why his results would be adjusted. In my case, I'm adjusting the result itself, in Silver's, he's giving it far less weight.

You can look at Silver's analysis before 2010, and he praises Rasmussen. Since then, I was picking up skew without knowing why, but it might be more.

My complaint is that Silver isn't challenging his own employer. The CBS/NYT polls seem so skewed I'm not even using them. I looked into them, and found a D adjustment that can't be defended (+9 last month when I looked). Silver called the CBS/NYT one of the worst in 2010, yet, now that he's paid by them, he's giving it his strongest weight.

This doesn't invalidate Silver's results or methodology. But it does call into question a bias.

My question to you and others is, given you have a point about Rasmussen, why are similar criticisms of polls that favor the Democrats defended so vigorously? Why can't this work both ways? Is your side always right, no matter what?

I think I have the answer to this. It looks like in 2012 NYTimes partnered with Quinnipiac to do their polling. Quinnipiac has an excellent reputation and methodology, hence their high weighting. In 2010 NYTimes was not partnered with Quinnipiac.

I think all of that is accurate, but I'm having a hard time finding confirmation.

[RainMaker]

Yes, they use Quinnipiac which was considerably accurate in the last election cycle. Solecismic is trying to compare a news publication (New York Times) to a polling agency (Rasmussen). He should be comparing Quinnipiac to Rasmussen.

[Solecismic]

Originally Posted by Passacaglia

But if I assume the errors on this score are distributed the way I prefer them to be -- huzzah!!

Seems that way, doesn't it? In a liberal crowd, Rasmussen could be polling by rescuing kittens from trees and he'd be blasted for it.

I've made a case, over and over. It's not like I even care - you take everything partisan with a grain of salt. Either way. But to a partisan, there is no such thing as either way. It's my way or you're biased.

Cell phones? How do you get a representative sample from cell phones? Don't know, don't care, don't want to know. But I do know that if Rasmussen doesn't use cell phones, the liberal side will blast him for that. And if he does use cell phones, then the liberal side will say no one responsible would use them.

I get tired of being called a Republican when I've never voted for one for president. And I don't want to waste any more time explaining why I think Rasmussen's data is valuable even if it has to be massaged (just like PPP's the other way).

I don't know who's right this time. My gut instinct is that Silver is. But I don't like how he's getting there.

[RainMaker]

Originally Posted by Solecismic

I've made a case, over and over. It's not like I even care - you take everything partisan with a grain of salt. Either way. But to a partisan, there is no such thing as either way. It's my way or you're biased.

You haven't made a case at all. You just say Silver is bias....just cuz.

[Solecismic]

Originally Posted by RainMaker

You haven't made a case at all. You just say Silver is bias....just cuz.

Go back and read then. Just cuz.

[DaddyTorgo]

Originally Posted by Solecismic

Seems that way, doesn't it? In a liberal crowd, Rasmussen could be polling by rescuing kittens from trees and he'd be blasted for it.

I've made a case, over and over. It's not like I even care - you take everything partisan with a grain of salt. Either way. But to a partisan, there is no such thing as either way. It's my way or you're biased.

Cell phones? How do you get a representative sample from cell phones? Don't know, don't care, don't want to know. But I do know that if Rasmussen doesn't use cell phones, the liberal side will blast him for that. And if he does use cell phones, then the liberal side will say no one responsible would use them.

I get tired of being called a Republican when I've never voted for one for president. And I don't want to waste any more time explaining why I think Rasmussen's data is valuable even if it has to be massaged (just like PPP's the other way).

I don't know who's right this time. My gut instinct is that Silver is. But I don't like how he's getting there.

Dude - weighting it differently is just another way of massaging it.

[RainMaker]

Originally Posted by Solecismic

Go back and read then. Just cuz.

I did.

Originally Posted by Solecismic

I don't know who's right this time. My gut instinct is that Silver is. But I don't like how he's getting there.

Gut instinct vs math.

[Solecismic]

Originally Posted by DaddyTorgo

Dude - weighting it differently is just another way of massaging it.

I think you're all missing the point here. Both sides need massaging. If Silver is giving full weight to a poll that produces outlying results and assumes a D+9 (as it did recently - not, apparently, its last poll), then he's falling prey to emotion (or an employer) here.

It's not that difficult a concept to grasp, especially since your guy is very direct about what he's doing. He believes Obama is being underpolled, so he weights accordingly. He may be right, he may be wrong.

But please, let's just agree to stop this "my guy is perfect, those who have the opposite view are charlatans" politics.

[Buccaneer]

Originally Posted by Solecismic

But please, let's just agree to stop this "my guy is perfect, those who have the opposite view are charlatans" politics.

In this crowd (esp. those that have been dominating this thread all year), hell would freeze over before that.

[sterlingice]

The goal for any pollster is to be correct, right? And if a polling agency is correct next week with their projections, their sampling bias is correct, too, right? That bias was the correct way to sample the polled individuals for this election.

Ultimately one has the make some assumptions on how to turn a 500 or 1000 person sample into representative of a 130M voting populace (as of 2008). If the results are correct next week, wasn't that the best way to sample (and there could be better, but of the ones out this time)?

So we'll know next week, who had the best sampling. There can't really be a "I think X is correct but their methodology is wrong" when the goal is to correctly guess the result.

SI

[Buccaneer]

SI, or just plain lucky.

[JonInMiddleGA]

Originally Posted by sterlingice

The goal for any pollster is to be correct, right?

Not necessarily.

The primary goal for most pollsters would almost certainly be to provide the client with what they're paying for. Sometimes that might not always be accurate results.

[sterlingice]

Originally Posted by JonInMiddleGA

Not necessarily.

The primary goal for most pollsters would almost certainly be to provide the client with what they're paying for. Sometimes that might not always be accurate results.

True. For some groups, there is a value to being biased one way or another to make the race appear closer or leaning one direction.

But isn't there a lot of money in being accurate, as well?

SI

[DaddyTorgo]

Originally Posted by sterlingice

True. For some groups, there is a value to being biased one way or another to make the race appear closer or leaning one direction.

But isn't there a lot of money in being accurate, as well?

SI

Particularly for someone who's publicly aggregating polls, not trying to create future business for themselves.

[JonInMiddleGA]

Originally Posted by sterlingice

True. For some groups, there is a value to being biased one way or another to make the race appear closer or leaning one direction. But isn't there a lot of money in being accurate, as well?

Theoretically yes. But, theoretically, if you're good enough you can be accurate on the polls where that's the primary goal & skew the data enough to provide the desired result on polls where that's the objective. It doesn't have to be either/or.

[JonInMiddleGA]

Originally Posted by DaddyTorgo

Particularly for someone who's publicly aggregating polls, not trying to create future business for themselves.

He's creating content for the New York Times, who signed him to a licensing deal in 2010. As long as he generates traffic, and/or whatever result they desire, he's doing the job he's been/being paid for.

[DaddyTorgo]

Originally Posted by JonInMiddleGA

He's creating content for the New York Times, who signed him to a licensing deal in 2010. As long as he generates traffic, and/or whatever result they desire, he's doing the job he's been/being paid for.

And if his results don't bear out or aren't at least plausible then he'll be a laughingstock and won't generate traffic.

He called the Republican wave in 2010, as has been noted here...not sure where this huge argument for him being partisan or his methodology being so suspect comes from - it has no basis in fact (given what we know right now about turnout in the future).

[JonInMiddleGA]

Originally Posted by DaddyTorgo

And if his results don't bear out or aren't at least plausible then he'll be a laughingstock and won't generate traffic.

The license expires in 2013 IIRC.

Noting, just FTR, that I haven't said anything about Silver's work one way or the other (besides the specific NYT fact). Mostly commenting on the reason(s) polling may not always have accuracy as the only/top priority.

[larrymcg421]

I'm sure glad Bucc came in here to tell us once again how much better he is than everyone else.

Look, Jim made a claim that Silver was biased towards Democrats because 1) He includes PPP 2) He includes CBS/NYT and 3) He has a lower weighting for Rasmussen than they deserve.

He's offered no evidence that PPP shouldn't be included or given a strong weight based on their past performance.

He claims that Silver gives CBS/NYT a higher weight than they deserve because he works for NYT, but provides no evidence to suggest they don't deserve that rating other than one poll a month ago that had a D+9 weighting.

He claims that Rasmussen is being given a lower weight than they deserve because Silver is partisan and has it out for a Republican leaning pollster. He provides evidence that they were right on the 2008 national race. I countered that by showing how badly they performed in the 2008 state polls and the 2010 Generic Congressional ballot (in which CBS/NYT called it within .8 pts).

Just because someone is a liberal and they weight a Democratic pollster higher than a Republican pollster doesn't automatically mean they are wrong. If Silver didn't provide his methodology and his weighting formulas, then sure he'd be suspect. If there was evidence that the Republican leaning pollster deserves a higher rating, then show it. But just arguing that Silver is a Democrat and thats his reason for downgrading Rasmussen is amateurish especially in light of evidence presented by people in this thread that shows how bad they've performed in the past.

[RainMaker]

And I don't care about the politics of it, I just don't like hatred toward math/science. I'd feel the same way if someone was arguing that someone is a great 3 point shooter while I was looking at his stats that showed he shot 20% from there. Then when asked why they felt that way, they say it's a gut instinct.

[Buccaneer]

Well, you do have to admit that no one (including me) is being (or should be) objective here. Every argument is seen through the prism of wanting your guy to win (or the other guy to lose). It's what being a partisan is all about, whether as a postive or negative force, or both. It's also like a sports team that you have emotionally invested in (and/or emotionally hating a rival team or its fans), you care about the outcome and want others to be persuaded. Nothing wrong with any of that and it has been going on for a long time here, including here at FOFC. Calling it like I see it.

[larrymcg421]

Rasmussen is 5 pts to the right of everyone in WI, 5 pts to the right of everyone in OH, 4 pts to the right of everyone in NH, and 4 pts to the right of everyone in CO. You don't see a pollster that consistently far to the left in several states and certainly not CBS/NYT or PPP.

It's not impossible that Rasmussen is just that much smarter than everyone else. If so, he'll be a huge star come Nov. 7th. Good luck to anyone betting on that.

[JPhillips]

Originally Posted by Solecismic

I think you're all missing the point here. Both sides need massaging. If Silver is giving full weight to a poll that produces outlying results and assumes a D+9 (as it did recently - not, apparently, its last poll), then he's falling prey to emotion (or an employer) here.

It's not that difficult a concept to grasp, especially since your guy is very direct about what he's doing. He believes Obama is being underpolled, so he weights accordingly. He may be right, he may be wrong.

But please, let's just agree to stop this "my guy is perfect, those who have the opposite view are charlatans" politics.

Weighting and adjustments for house effects are two different processes. If Quinnipiac is consistently more Dem than others that will be taken into consideration in the model. Weighting is based on past performance and methodology and generally doesn't change within an election cycle. House effect is measured constantly and adjusted as it changes. (Right now Rasmussen's house effect isn't even two points. ) - edit Maybe not true now. That number was from August.

One polls composition, though, doesn't mean shit.

[DaddyTorgo]

Originally Posted by Buccaneer

Well, you do have to admit that no one (including me) is being (or should be) objective here. Every argument is seen through the prism of wanting your guy to win (or the other guy to lose).

Disagree.

[molson]

So it's just a coincidence then that all the liberal posters think the pro-liberal predictions are right and all the conservative posters think the pro-conservative predictions are right?

[JPhillips]

Originally Posted by Buccaneer

Well, you do have to admit that no one (including me) is being (or should be) objective here. Every argument is seen through the prism of wanting your guy to win (or the other guy to lose). It's what being a partisan is all about, whether as a postive or negative force, or both. It's also like a sports team that you have emotionally invested in (and/or emotionally hating a rival team or its fans), you care about the outcome and want others to be persuaded. Nothing wrong with any of that and it has been going on for a long time here, including here at FOFC. Calling it like I see it.

But in terms of Silver I don't care who he has winning. I want as accurate information as possible. Winning the 538 prediction doesn't mean jack to me, especially if I suspect it's wrong.

Again, Silver's model may not be perfect, but he's transparent in his methodology and has no reason to be biased towards inaccurate information. The people who say otherwise should provide more proof than the fact that he works for the NYT. I hate that we've gotten to a point where all information is suspect. I read a number of conservatives and value their information, I think Dan Larison is one of the smartest foreign policies guys out there. I don't immediately discount anyone that votes GOP.

[JPhillips]

Originally Posted by molson

So it's just a coincidence then that all the liberal posters think the pro-liberal predictions are right and all the conservative posters think the pro-conservative predictions are right?

I've specifically said that I don't know if Silver is right, but I am much more certain that he isn't showing a political bias. He has a good track record, but every election is different and the data isn't extensive. He may be off this year, but he's at least trying to develop a prediction model based on data.

[Warhammer]

Originally Posted by larrymcg421

Rasmussen is 5 pts to the right of everyone in WI, 5 pts to the right of everyone in OH, 4 pts to the right of everyone in NH, and 4 pts to the right of everyone in CO. You don't see a pollster that consistently far to the left in several states and certainly not CBS/NYT or PPP.

It's not impossible that Rasmussen is just that much smarter than everyone else. If so, he'll be a huge star come Nov. 7th. Good luck to anyone betting on that.

Gallup is actually to the right of Rasmussen on the national polls. I believe they are both approx. R+2/3.

As I mentioned previously, it is going to be very interesting to see how these work out. CBS/Quinnipac and PPP have been D+8 for many polls which is based upon the 2008 Election turnout. That was a historic high, and certainly the biggest margin in party turnout in the past 30 years. Relying on that sort of turnout again (combined with an abyssmal R turnout in 2008) is ridiculous. The Gallup and Rasmussen polls are in line with the 2004 election which was a high turnout election which was R+4.

The message is CBS/Quinnipac and PPP are banking on things being identical to 2008 (which I have serious questions about) whereas Gallup and Rasmussen are banking on something similar to 2004, but not as pronounced (which I think is more likely). The race is so flipping close though that the skew of the poll has a huge impact on who is winning where.

[sterlingice]

Originally Posted by molson

So it's just a coincidence then that all the liberal posters think the pro-liberal predictions are right and all the conservative posters think the pro-conservative predictions are right?

This isn't like policy where there are many ways to solve a lot of problems. Next Tuesday, one poll or set of polls is going to show to be more accurate than the others unless they all converge to very similar conclusions (at which point, I would worry about the non-transparent ones just making some quick and dirty house adjustments) or they are all equally bad.

If Romney wins by +3, Rasmussen will probably be the most accurate as long as there's not some oddball state orientation. If Obama wins by +3, then it's probably PPP that will look the best, again, barring some oddball orientation. Now, no poll is going to be perfect across all 50 states but that plus the Senate races gives a pretty large sample size each election to weight against (~80 odd races). And, yes, finding who is most correct does show who is best at predicting and most accurate not "who is most biased" or "lucky".

Someone will be correct next week. And I really think chalking it up to "luck" or saying "well, your side (whichever side that is) just was right because you wanted it to be that way" is just patently false. If PPP is oversampling Democrats, they will show up as being 3% off in every race, even if Obama wins. And if Rasmussen is undersampling Democrats and it causes them to be 3% off, they will show up at being 3% off in every race, even if Romney wins.

And we can show next week who is most accurate.

SI

[larrymcg421]

Originally Posted by molson

So it's just a coincidence then that all the liberal posters think the pro-liberal predictions are right and all the conservative posters think the pro-conservative predictions are right?

Not arguing that it's a coincidence, but just because that difference exists doesn't mean that both sides are wrong or looking at the information in a skewed manner. More of the predictions favor the liberal argument than the conservative argument at this point. Maybe the conservative pollsters are just better, but I've provided quite a bit of evidence for why I don't think that's the case. I'm not arrogant enough to think there isn't a valid response to my claims, but "you're liberal and favoring a liberal prediction" isn't a valid response to a statistical argument.

[Warhammer]

The other item that I think is telling, why is Obama spending money in MN? If Romney is getting close in MN, of all places, Obama is in deep trouble.

[Warhammer]

Originally Posted by larrymcg421

Not arguing that it's a coincidence, but just because that difference exists doesn't mean that both sides are wrong or looking at the information in a skewed manner. More of the predictions favor the liberal argument than the conservative argument at this point. Maybe the conservative pollsters are just better, but I've provided quite a bit of evidence for why I don't think that's the case. I'm not arrogant enough to think there isn't a valid response to my claims, but "you're liberal and favoring a liberal prediction" isn't a valid response to a statistical argument.

Again, methodology will be very interesting to me. Many polls (as I pointed out) are using a D+8 skew for their results. They are basing this on the 2008 election.

My point, from what I have seen anecdotally, is this is the wrong way to look at this election. Unless you think R turnout, is going to be historically low again, and the D turnout will be high again, there is no way to make the D+8 model to hold up.

Now, is the R+2/3 the correct model? I'm not sure, but I think that is a heck of a lot more likely this election as fired up as the right is, then the D+8 model. With the election as close as it is at D+8, I think it is extremely bad news for Obama.

[molson]

Originally Posted by sterlingice

And we can show next week who is most accurate.

Well, I'm sure the winners will claim that about whatever poll they supported, but the polls are just offering probabilities, right, with a sample size of 1 presidential election? I don't think we can ever know for sure whose polls are the best. That makes it a irresistible argument, kind of, because everyone can always claim they're right. I don't think the Dems will be burying Silver here if Romney wins, they'll just emphasize the "probability" aspect of this (and maybe talk about the Hurricane and voter ID laws or something). If Obama wins though, even if it's razor thin, I'm sure that will be presented as proof of the superiority of the polls that went that way. And similar justifications can be made from the other side whichever way things swing. I mean, this thing is close enough that I just hope we have a clear election winner, I don't think there's any chance that there's a clear correct poll at the end of this.

[larrymcg421]

Originally Posted by Warhammer

Gallup is actually to the right of Rasmussen on the national polls. I believe they are both approx. R+2/3.

As I mentioned previously, it is going to be very interesting to see how these work out. CBS/Quinnipac and PPP have been D+8 for many polls which is based upon the 2008 Election turnout. That was a historic high, and certainly the biggest margin in party turnout in the past 30 years. Relying on that sort of turnout again (combined with an abyssmal R turnout in 2008) is ridiculous. The Gallup and Rasmussen polls are in line with the 2004 election which was a high turnout election which was R+4.

The message is CBS/Quinnipac and PPP are banking on things being identical to 2008 (which I have serious questions about) whereas Gallup and Rasmussen are banking on something similar to 2004, but not as pronounced (which I think is more likely). The race is so flipping close though that the skew of the poll has a huge impact on who is winning where.

I've already refuted this. The last CBS/NYT poll had a D+5 weighting and the last PPP poll had a D+2 weighting. None of the current polls shows a D+8.

R+4 is absurd. That would be 4 pts more Republican than two years ago, which was one of the most toxic environments for Democrats ever.