Kilmer17s non confrontational election prediction-

[nonniey]

I look at the "no enthusiasm for Obama" narrative a little skeptical-ish.

I think it's only natural that after 4 years in office, and the country still not quite "out of the woods yet" there isn't going to be the same buzz in the air that there was in 2008

It can't ever be like 2008 again, for many reasons, but that doesn't mean that Obama voters are not going to turn out in 2012, it might just simply mean that after 4 years of Obama being on the TV screen,and in our homes on a daily basis, the "aura" about him is gone.

Lets not forget, before 2008, a lot of us had known him mostly for his speech at the 2004 DNC, so there was something new and fresh about him as a Presidential candidate.

Whether you are voting for Obama enthusiastically, or because you don't like what you've seen from his potential replacement, it is still a vote for Obama, and likewise with Romney.

Maybe but the big assumption from the Obama campaign and many pollsters is that the Romney supporters are going to replicate the McCain turnout, in addition to assuming Obama has not only not lost any support he has increased support over 2008.

[Bliz]

Maybe but the big assumption from the Obama campaign and many pollsters is that the Romney supporters are going to replicate the McCain turnout, in addition to assuming Obama has not only not lost any support he has increased support over 2008.

No one, inside the campaign or out, is assuming that. Everyone...pollsters, run of the mill dems, people inside the campaign, Obama himself...literally everyone recognizes that this will be a tighter race and he won't have as much support as he had in 2008.

This idea about those assumptions is a narrative for the conservative media to attack polls that don't say what they want the polls to say. That's about it. Not a shred of truth or reason to it.

[Larry]

Then explain polls that have Democrats with plus 10 in their polling samples when the final difference in 2008 was plus 7.

It's already been explained. In this thread. In fact, on this page of this thread.

And in several others.

But it doesn't fit the Republican agenda. So you ignore it.

Now explain why nobody at ES is capable of setting the clocks on the servers.

[nonniey]

No one, inside the campaign or out, is assuming that. Everyone...pollsters, run of the mill dems, people inside the campaign, Obama himself...literally everyone recognizes that this will be a tighter race and he won't have as much support as he had in 2008.

This idea about those assumptions is a narrative for the conservative media to attack polls that don't say what they want the polls to say. That's about it. Not a shred of truth or reason to it.

Then explain polls that have Democrats with plus 10 in their polling samples when the final difference in 2008 was plus 7. As I pointed out earlier in Ohio for example the three polls that have Obama slightly ahead use polls where the Democrats match or exceed Democrat participation in 2008. Axelrod , MSNBC and others are basing their prediction of victory on these polls.

I'm not saying they are wrong I was just asking the board if they believe the Democrats will outperform their 2008 performance vice the Republicans (of plus 7). Larry answered in the affirmative (The only one so far) but you and many others seem to be saying no (In which case confidence in an Obama victory should be a little less on this board than it is).

[nonniey]

It is my understanding that this is not at all how it works. As I understand it, the polls call a lot of people and ASK them what their party affiliation is during the phone call. If they are getting 5 percent more people who say they are Democrats, that is what the poll will reflect. If they get 5 percent more that say they are GOP, then that is what the poll will reflect. The only adjusting they do is demographic, i.e., the older people are, the more likely they are to actually vote rather than to just say they are going to vote.

I think I may have tracked down how the Pollsters establish their models. Looking at the Quinnipac poll released today the poll has the President up by 5 in Ohio, 1 in Virginia, and 1 in Florida. I'm sure that will please many on this board, but again it goes back to an assumption for this poll that the Democrats will exceed their performance by 33-50% over 2008. How they meet this assumption is by weighting the responses given party affiliation which tells me what they are reporting is not the actual raw percentage of the responses (ie more people could have actually responded for Romney in these polls but when weighted that dropped the over all percentage reported).


See below

Florida

LIKELY VOTERS........
Weighted UnWeighted
Percent Frequency
PARTY IDENTIFICATION
Republican 30% 352
Democrat 37 350
Independent 29 333
Other/DK/NA 4 38




Virginia
LIKELY VOTERS........
Weighted UnWeighted
Percent Frequency
PARTY IDENTIFICATION
Republican 27% 290
Democrat 35 349
Independent 35 395
Other/DK/NA 3 40

Ohio

LIKELY VOTERS........
Weighted UnWeighted
Percent Frequency
PARTY IDENTIFICATION
Republican 29% 343
Democrat 37 396
Independent 30 334
Other/DK/NA 4 37

---------- Post added October-31st-2012 at 06:42 PM ----------

It's already been explained. In this thread. In fact, on this page of this thread.

And in several others.

But it doesn't fit the Republican agenda. So you ignore it.

Now explain why nobody at ES is capable of setting the clocks on the servers.

Turns out those explanations were apparently wrong. At least one pollster (Quinnipac) weights its' responses to meet the assumption that the Democrats will outperform their 2008 performance (I suspect all the major pollsters wieght resposnes to meet a model). You in fact are the only one so far in this forum that believes the assumption is valid which is ok and fully supports a prediction of an Obama victory, it also mirrors the Obama campaign belief. However those in this forum that are saying they don't believe those assumptions are correct are the ones whose optimism is a bit baseless.

[Predicto]

I think I may have tracked down how the Pollsters establish their models. Looking at the Quinnipac poll released today the poll has the President up by 5 in Ohio, 1 in Virginia, and 1 in Florida. I'm sure that will please many on this board, but again it goes back to an assumption for this poll that the Democrats will exceed their performance by 33-50% over 2008. How they meet this assumption is by weighting the responses given party affiliation which tells me what they are reporting is not the actual raw percentage of the responses (ie more people could have actually responded for Romney in these polls but when weighted that dropped the over all percentage reported).


See below

Florida

LIKELY VOTERS........
Weighted UnWeighted
Percent Frequency
PARTY IDENTIFICATION
Republican 30% 352
Democrat 37 350
Independent 29 333
Other/DK/NA 4 38




Virginia
LIKELY VOTERS........
Weighted UnWeighted
Percent Frequency
PARTY IDENTIFICATION
Republican 27% 290
Democrat 35 349
Independent 35 395
Other/DK/NA 3 40

Ohio

LIKELY VOTERS........
Weighted UnWeighted
Percent Frequency
PARTY IDENTIFICATION
Republican 29% 343
Democrat 37 396
Independent 30 334
Other/DK/NA 4 37

---------- Post added October-31st-2012 at 06:42 PM ----------



Turns out those explanations were apparently wrong. At least one pollster (Quinnipac) weights its' responses to meet the assumption that the Democrats will outperform their 2008 performance (I suspect all the major pollsters wieght resposnes to meet a model). You in fact are the only one so far in this forum that believes the assumption is valid which is ok and fully supports a prediction of an Obama victory, it also mirrors the Obama campaign belief. However those in this forum that are saying they don't believe those assumptions are correct are the ones whose optimism is a bit baseless.

I honestly believe you are misunderstanding this, but there is no point in belaboring it any further.

[nonniey]

I honestly believe you are misunderstanding this, but there is no point in belaboring it any further.

It doesn't seem that I'm misunderstanding this.

Here is how weighting is explained by PrecisionPolling.com: "It is frequently the case that the people who answered your poll are not fully representative of the region you were polling over. Weighting is a technique to adjust answers to account for over- and under-represented groups."

I'm going to wait to see how Sandy plays out but if everything remains the same by Monday I'll predict easy Romney victories in at least Florida and VA and a close Romney Victory in Ohio. That is based on my belief that Democrats will not come close to matching the turnout advantage they had over Republicans they had in 2008. It would not surprise me if the Republicnas end up with the advantage this time around (plus 2 or 3) and if the latter happens it will be a landslide for Romney..

[Predicto]

It doesn't seem that I'm misunderstanding this.

Here is how weighting is explained by PrecisionPolling.com: "It is frequently the case that the people who answered your poll are not fully representative of the region you were polling over. Weighting is a technique to adjust answers to account for over- and under-represented groups."

I'm going to wait to see how Sandy plays out but if everything remains the same by Monday I'll predict easy Romney victories in at least Florida and VA and a close Romney Victory in Ohio. That is based on my belief that Democrats will not come close to matching the turnout advantage they had over Republicans they had in 2008. It would not surprise me if the Republicnas end up with the advantage this time around (plus 2 or 3) and if the latter happens it will be a landslide for Romney..

I know what weighting is.

But virtually every major pollster says that they do not do any weighting at all to adjust for party affiliation. They weigh for demographics but they don't change their results to change the republican/democratic mix that they get from their responses. If they get plus 8 Percent GOP responses, that's what they report. That's because party identification is one of the things that polling is supposed to be measuring - it changes from election to election.

Here's Pew talking about it.
http://www.people-press.org/2012/08/...lection-polls/

[twa]

Predicto, do you think they adjust the sampling?
different terms give the same result

both weighed and unweighed have value, but you have to look beyond the surface

[nonniey]

I know what weighting is.

But virtually every major pollster says that they do not do any weighting at all to adjust for party affiliation. They weigh for demographics but they don't change their results to change the republican/democratic mix that they get from their responses. If they get plus 8 Percent GOP responses, that's what they report. That's because party identification is one of the things that polling is supposed to be measuring - it changes from election to election.

Here's Pew talking about it.
http://www.people-press.org/2012/08/...lection-polls/

Did you not see what I posted from the poll itself?? That is exactly what Quinnipac did. Look at the Florida example again. They show the raw numbers and more Republicans than Democrats were interviewed (352 Rs, 350 Dems) but the poll ended with a 37 to 30 Democratic advantage . This is because they assess that is what the final turnout will be. It is an assumption and pretty sure most polls do this,. Quinnipac could be making the right assumption but is is still an assumption, Rassmussen for example has a different assumption and gives a plus 2 advantage to the Dems (he was interviewed and he mentioned this tonight). The question is who is correct. Gallup weighs their national projection giving the Republicans a plus 1 advantage. Bottomline party affiliation is in facted weighed to fit the model.


Florida
LIKELY VOTERS........
Weighted UnWeighted
Percent Frequency
PARTY IDENTIFICATION
Republican 30% 352
Democrat 37 350
Independent 29 333
Other/DK/NA 4 38

[Predicto]

Predicto, do you think they adjust the sampling?
different terms give the same result

both weighed and unweighed have value, but you have to look beyond the surface

Yes, they adjust the sampling. If they reach only 20 percent senior citizens, but the district demographics and history show that 33 percent of the voters will be senior citizens, then they adjust the results to reflect this.

But if they get 40 percent self identified Democrats and 43 percent GOP, they don't adjust based on that. If they did, then what would they be measuring? If more voters are identifying as GOP than last time, that is because the GOP is more popular than last time. That's why people hold elections in the first place.

The idea that polls adjust to add more Democrats or Republicans is a myth.

---------- Post added October-31st-2012 at 09:14 PM ----------

Did you not see what I posted from the poll itself?? That is exactly what Quinnipac did. See posts above.

No it isn't. It is what you think they did.

They adjusted based on demographics, and (in that example) the results changed in favor of the Democrats. In other cases it goes the other way.

Imagine a district that is 50 percent women, but for some reason your phone calls on this poll happened to reach 80 percent men and only 20 percent women. That result would be adjusted - and a side effect would be an increase in the predicted Democratic vote percentage. But it could go the other way the next poll, when you managed to call 80 percent women and only 20 percent men.

Does that make sense?

[Duckus]

Did you not see what I posted from the poll itself?? That is exactly what Quinnipac did. Look at the Florida example again. They show the raw numbers and more Republicans than Democrats were interviewed (352 Rs, 350 Dems) but the poll ended with a 37 to 30 Democratic advantage . This is because they assess that is what the final turnout will be. It is an assumption and pretty sure most polls do this,. Quinnipac could be making the right assumption but is is still an assumption, Rassmussen for example has a different assumption and gives a plus 2 advantage to the Dems (he was interviewed and he mentioned this tonight). The question is who is correct. Gallup weighs their national projection giving the Republicans a plus 1 advantage. Bottomline party affiliation is in facted weighed to fit the model.


Florida
LIKELY VOTERS........
Weighted UnWeighted
Percent Frequency
PARTY IDENTIFICATION
Republican 30% 352
Democrat 37 350
Independent 29 333
Other/DK/NA 4 38

You are confusing demographics with party ID. This is probably the biggest and most common misunderstanding of polling.

"If a pollster weights by party ID, they are substituting their own judgment as to what the electorate is going to look like. It's not scientific," said Doug Schwartz, the director of the Quinnipiac University Polling Institute, which doesn't weight its surveys by party identification.

http://m.theatlantic.com/politics/ar...licans/262834/

[Veretax]

if you guys want some opinion on why the polls, many of them are oversampling, over biasing, and over adjusting in favor of D's check out Dick Morris's latest blog

http://www.dickmorris.com/why-the-ny...aign=dmreports

Now, all of this is really speculation, based on the opinions of the people doing 'adjustments' and such We all know how split the country has felt in the last 20 years.

[LadySkinsFan]

For an astrological look at the election, I give you these two articles. Interesting reads, and only to be taken seriously if you want to.

First is about the election, Mercury turns retrograde on election day.

http://planetwaves.net/astrologynews...ction-day.html

The second is about Obama's and Romney's charts. Warning: in the opening, the author incorrectly names Romney a couple of times, which I think is disrespectful, and the author also gives a disclaimer in the 3rd paragraph.

http://planetwaves.net/astrologynews/1770564627.html

[JimboDaMan]

if you guys want some opinion on why the polls, many of them are oversampling, over biasing, and over adjusting in favor of D's check out Dick Morris's latest blog

http://www.dickmorris.com/why-the-ny...aign=dmreports

Now, all of this is really speculation, based on the opinions of the people doing 'adjustments' and such We all know how split the country has felt in the last 20 years.

I don't understand. Are you (and Dick) saying the pollsters are lying, that they claim to not adjust for party affiliation when in fact they are? I get why Dick would claim this, because he's pushing a certain agenda, and because he's, well, a Dick. But seriously, does it make sense that the pollsters who's livelihood depends on accuracy would actually lie, and would deliberately publish results they know will eventually be shown to be inaccurate?

In the end it may turn out that Dick's prediction is closer to the mark then most pollsters. If so, the pollsters will clearly have some adjustments to make to their methodology. But the reason will not be because Dick adjusted for expected party affiliation and they did not (or as he seems to be claiming, did adjust but did so incorrectly).