According to the swing state polls Obama is set to win a second term

[Foxfyre]

RCP has Obama ahead in all battleground except for Florida and North Carolina.

Over-sampling will lead to silly, useless, irrelevant skewed results.

At any rate, the only poll that ultimately counts is the vote tomorrow. I have been watching all the major polls for some time now and it will be interesting to see which ones have been the most accurate. For some time now Rasmussen has had an excellent track record but no way he's attempting to call the election anywhere at this time in any of the close ones.

 

[Liability]

UnSkewed Polls -- erasing the bias to show an accurate picture of politics


Yeah yeah. Just another prognostication.

But it may prove to be closer to the final outcome than all the rest.

 

[AceRothstein]

UnSkewed Polls -- erasing the bias to show an accurate picture of politics


Yeah yeah. Just another prognostication.

But it may prove to be closer to the final outcome than all the rest.

That is the first unskewedpolls map I've seen that MAY actually end up happening. The rest have been pure fantasy.

 

[theDoctorisIn]

Party identification is often the most important demographnic in poll weighting when the polls are to evaluate whether one candidate is ahead of another.

http://courses.ttu.edu/hdfs3390-reifman/weighting.htm

The demographics are divided into those who identify as Republicans and Democrats with everybody else thrown into the Independents division.

Did you read your own link? It basically argues against you and shows that only 2 pollsters currently weight by Party ID. It is most definitely not the industry standard to do so.

That depends on what industry you are referring to. Scientific pollsters will strive for as accurate a tally as possible. Media polls are too often looking for advantage for their preferred candidate and the alphabet channels are pretty well in bed with Obama. You seem to be looking for arguments weighted in favor of your point of view so you won't have to admit you didn't know what you were talking about. And so it goes. The campaigns themselves conduct their owninternal polls which are likely to be the most accurate out there but they are never published.

How about providing a link from a reliable sourvce stating that the industry standard is not to address party affiliation as you have stated? Good luck with that though.

You already provided that link...

Media polls are almost always contracted out to "scientific poll" companies. NBC doesn't do their own polling, they hire a company to do it. And no, they're not looking to give an "advantage" to any candidates.

If it makes you feel better to cling to the hope that all the polls are wrong, feel free to do so.

 

[Foxfyre]

Did you read your own link? It basically argues against you and shows that only 2 pollsters currently weight by Party ID. It is most definitely not the industry standard to do so.

That depends on what industry you are referring to. Scientific pollsters will strive for as accurate a tally as possible. Media polls are too often looking for advantage for their preferred candidate and the alphabet channels are pretty well in bed with Obama. You seem to be looking for arguments weighted in favor of your point of view so you won't have to admit you didn't know what you were talking about. And so it goes. The campaigns themselves conduct their owninternal polls which are likely to be the most accurate out there but they are never published.

How about providing a link from a reliable sourvce stating that the industry standard is not to address party affiliation as you have stated? Good luck with that though.

Why weighting polls for party identification is wishful thinking | Harry J Enten | Comment is free | guardian.co.uk

Survey Says

The Morning Plum: No, pollsters are not conspiring to destroy Romney - The Plum Line - The Washington Post

Dick Morris: Party Disparities Aren

The Debate Over Party ID in Political Polling - The Numbers Guy - WSJ

Hopefully these are enough links for you and reliable enough sources for you.

Your assertion has been that the polls aren't weighted by party identification. All three of your links assume that they are and why weighting by party identification is often an inaccurate way to do it. So you are now agreeing with me that the polls are often inaccurately weighted by party identification?

 

[AceRothstein]

That depends on what industry you are referring to. Scientific pollsters will strive for as accurate a tally as possible. Media polls are too often looking for advantage for their preferred candidate and the alphabet channels are pretty well in bed with Obama. You seem to be looking for arguments weighted in favor of your point of view so you won't have to admit you didn't know what you were talking about. And so it goes. The campaigns themselves conduct their owninternal polls which are likely to be the most accurate out there but they are never published.

How about providing a link from a reliable sourvce stating that the industry standard is not to address party affiliation as you have stated? Good luck with that though.

Why weighting polls for party identification is wishful thinking | Harry J Enten | Comment is free | guardian.co.uk

Survey Says

The Morning Plum: No, pollsters are not conspiring to destroy Romney - The Plum Line - The Washington Post

Dick Morris: Party Disparities Aren

The Debate Over Party ID in Political Polling - The Numbers Guy - WSJ

Hopefully these are enough links for you and reliable enough sources for you.

Your assertion has been that the polls aren't weighted by party identification. All three of your links assume that they are and why weighting by party identification is often an inaccurate way to do it. So you are now agreeing with me that the polls are often inaccurately weighted by party identification?

Are you just playing dumb now? There are 5 links and none of them assume that the polls they are discussing are weighted by party ID. They are saying they are not weighted by party ID and that they are weighted by demographics such as age and race.

 

[amrchaos]

Well, "Happy kick Bronco Bama out of office day" eve!

 

[Foxfyre]

Why weighting polls for party identification is wishful thinking | Harry J Enten | Comment is free | guardian.co.uk

Survey Says

The Morning Plum: No, pollsters are not conspiring to destroy Romney - The Plum Line - The Washington Post

Dick Morris: Party Disparities Aren

The Debate Over Party ID in Political Polling - The Numbers Guy - WSJ

Hopefully these are enough links for you and reliable enough sources for you.

Your assertion has been that the polls aren't weighted by party identification. All three of your links assume that they are and why weighting by party identification is often an inaccurate way to do it. So you are now agreeing with me that the polls are often inaccurately weighted by party identification?

Are you just playing dumb now? There are 5 links and none of them assume that the polls they are discussing are weighted by party ID. They are saying they are not weighted by party ID and that they are weighted by demographics such as age and race.

Trust me. I am not 'dumb' on this subject. You obviously have not read my links or your links with any ability to understand what they are saying nor are you considering even what I have said. And you apparently do not know what 'weighting' means in regard to a scientific poll. It is far more intricate than just counting the number of Democrats and Republicans polled, but no poll can even hope to have any credibility at all without some idea of how the respondant has voted in the past or his/her party affiliation.

 

[AceRothstein]

Your assertion has been that the polls aren't weighted by party identification. All three of your links assume that they are and why weighting by party identification is often an inaccurate way to do it. So you are now agreeing with me that the polls are often inaccurately weighted by party identification?

Are you just playing dumb now? There are 5 links and none of them assume that the polls they are discussing are weighted by party ID. They are saying they are not weighted by party ID and that they are weighted by demographics such as age and race.

Trust me. I am not 'dumb' on this subject. You obviously have not read my links or your links with any ability to understand what they are saying nor are you considering even what I have said. And you apparently do not know what 'weighting' means in regard to a scientific poll. It is far more intricate than just counting the number of Democrats and Republicans polled, but no poll can even hope to have any credibility at all without some idea of how the respondant has voted in the past or his/her party affiliation.

I've read all links and none of them support your assertion that the polls are weighted by party ID or previous election turnout, not one of them. Well, except for Rasmussen.

 

[Foxfyre]

Are you just playing dumb now? There are 5 links and none of them assume that the polls they are discussing are weighted by party ID. They are saying they are not weighted by party ID and that they are weighted by demographics such as age and race.

Trust me. I am not 'dumb' on this subject. You obviously have not read my links or your links with any ability to understand what they are saying nor are you considering even what I have said. And you apparently do not know what 'weighting' means in regard to a scientific poll. It is far more intricate than just counting the number of Democrats and Republicans polled, but no poll can even hope to have any credibility at all without some idea of how the respondant has voted in the past or his/her party affiliation.

I've read all links and none of them support your assertion that the polls are weighted by party ID or previous election turnout, not one of them. Well, except for Rasmussen.

I really REALLY believe you haven't read any of them. Not even your own. But oh well. Somebody has to vote for Obama.

 

[AceRothstein]

Trust me. I am not 'dumb' on this subject. You obviously have not read my links or your links with any ability to understand what they are saying nor are you considering even what I have said. And you apparently do not know what 'weighting' means in regard to a scientific poll. It is far more intricate than just counting the number of Democrats and Republicans polled, but no poll can even hope to have any credibility at all without some idea of how the respondant has voted in the past or his/her party affiliation.

I've read all links and none of them support your assertion that the polls are weighted by party ID or previous election turnout, not one of them. Well, except for Rasmussen.

I really REALLY believe you haven't read any of them. Not even your own. But oh well. Somebody has to vote for Obama.

I HAVE read them. I don't understand how you can continue to say that they backup your argument. YOUR link didn't even do that.

Direct from the Quinnipiac page: "It is important to understand that all of the major polls – Gallup, Pew, Quinnipiac, CBS, ABC, and NBC- do not weight their data by party. Those polls, as does Quinnipiac, weight their data for things such as gender and age to match the Census data because these things don’t change."

 

[Foxfyre]

For those who want to know more about how polls are weighted, here is an excellent article explaining it. It is a UK writer, but the same criteria applies here:

UK Polling Report

 

[Uncensored2008]

 

[Foxfyre]

I've read all links and none of them support your assertion that the polls are weighted by party ID or previous election turnout, not one of them. Well, except for Rasmussen.

I really REALLY believe you haven't read any of them. Not even your own. But oh well. Somebody has to vote for Obama.

I HAVE read them. I don't understand how you can continue to say that they backup your argument. YOUR link didn't even do that.

Direct from the Quinnipiac page: "It is important to understand that all of the major polls – Gallup, Pew, Quinnipiac, CBS, ABC, and NBC- do not weight their data by party. Those polls, as does Quinnipiac, weight their data for things such as gender and age to match the Census data because these things don’t change."

They don't weight them by party alone. Or shouldn't. But in rebuttal that they don't weight them by party at all.....including Quinnipac....:

Election Polling 101 « Survey Says

 

[AceRothstein]

I really REALLY believe you haven't read any of them. Not even your own. But oh well. Somebody has to vote for Obama.

I HAVE read them. I don't understand how you can continue to say that they backup your argument. YOUR link didn't even do that.

Direct from the Quinnipiac page: "It is important to understand that all of the major polls – Gallup, Pew, Quinnipiac, CBS, ABC, and NBC- do not weight their data by party. Those polls, as does Quinnipiac, weight their data for things such as gender and age to match the Census data because these things don’t change."

They don't weight them by party alone. Or shouldn't. But in rebuttal that they don't weight them by party at all.....including Quinnipac....:

Election Polling 101 « Survey Says

It says right in my link THEY DO NO WEIGHT THEIR DATA BY PARTY. It says nothing once again about weighting by party.

 

[Pasco08]

On the Eve of the election they are neck and neck

 

[Foxfyre]

I HAVE read them. I don't understand how you can continue to say that they backup your argument. YOUR link didn't even do that.

Direct from the Quinnipiac page: "It is important to understand that all of the major polls – Gallup, Pew, Quinnipiac, CBS, ABC, and NBC- do not weight their data by party. Those polls, as does Quinnipiac, weight their data for things such as gender and age to match the Census data because these things don’t change."

They don't weight them by party alone. Or shouldn't. But in rebuttal that they don't weight them by party at all.....including Quinnipac....:

Election Polling 101 « Survey Says

It says right in my link THEY DO NO WEIGHT THEIR DATA BY PARTY. It says nothing once again about weighting by party.

If you can read that article and continue to assert that polls, including Quinnipac, are not weighted by political party then may I please suggest a good remedial reading course? What does it say in the very first paragraph? Never mind. You want to believe what you want to believe. There is simply no way in hell to do a political poll that involves all the demographics without including political party in the mix. But carry on. Again somebody has to vote for Obama. Might as well be you.

They may not vote according to party ID or registration, but to say they don't weight their polls by party identification is absurd when you look at the breakdown they use for the polls. How in the world could you evaluate a political race without identifying the political affiliations or identifications of those responding to the poll?

Here's another analysis:

http://www.redstate.com/realquiet/2...ng-hits-a-new-level-in-loaded-agenda-polling/

 

[AceRothstein]

They don't weight them by party alone. Or shouldn't. But in rebuttal that they don't weight them by party at all.....including Quinnipac....:

Election Polling 101 « Survey Says

It says right in my link THEY DO NO WEIGHT THEIR DATA BY PARTY. It says nothing once again about weighting by party.

If you can read that article and continue to assert that polls, including Quinnipac, are not weighted by political party then may I please suggest a good remedial reading course? What does it say in the very first paragraph? Never mind. You want to believe what you want to believe. There is simply no way in hell to do a political poll that involves all the demographics without including political party in the mix. But carry on. Again somebody has to vote for Obama. Might as well be you.

Are you dense? THIS LINE IS DIRECTLY FROM THE ARTICLE: "It is important to understand that all of the major polls – Gallup, Pew, Quinnipiac, CBS, ABC, and NBC- do not weight their data by party. Those polls, as does Quinnipiac, weight their data for things such as gender and age to match the Census data because these things don’t change."

And I'm the one that believes what I want to believe? You're either playing dumb or you have a reading comprehension problem.

 

[OODA_Loop]

Are you dense? THIS LINE IS DIRECTLY FROM THE ARTICLE: "It is important to understand that all of the major polls – Gallup, Pew, Quinnipiac, CBS, ABC, and NBC- do not weight their data by party. Those polls, as does Quinnipiac, weight their data for things such as gender and age to match the Census data because these things don’t change."

Post an example of one of those polls weighted even or for republicans and show everyone.

 

[Katzndogz]

I've read all links and none of them support your assertion that the polls are weighted by party ID or previous election turnout, not one of them. Well, except for Rasmussen.

I really REALLY believe you haven't read any of them. Not even your own. But oh well. Somebody has to vote for Obama.

I HAVE read them. I don't understand how you can continue to say that they backup your argument. YOUR link didn't even do that.

Direct from the Quinnipiac page: "It is important to understand that all of the major polls – Gallup, Pew, Quinnipiac, CBS, ABC, and NBC- do not weight their data by party. Those polls, as does Quinnipiac, weight their data for things such as gender and age to match the Census data because these things don’t change."

Weighting by turnout models doesn't mean weighting by party, even though it turns out that way most of the time. As was correctly pointed out, it's gender, age, race a number of factors based on who voted in the election year that they chose. In this case, it's 2008. When Cruz Bustamante was running against Arnold Schwarzeneggar for governor of California, the modeling was done by ethnicity and a mexican running for governor. They didn't weight by party because there are too few republicans in California to use in modeling. The polls had Bustamante ahead by double digits right before the vote and it ended up a landslide for Schwarzeneggar. The modeling was wrong. Had they used simply the raw data, it was apparent that Schwarzeneggar was going to win and win with the kind of margin that he had.

In today's polls we have no idea what the raw data says. That's a closely guarded secret among polling agencies. All we know is what we are told using the modeling.