• Welcome to OGBoards 10.0, keep in mind that we will be making LOTS of changes to smooth out the experience here and make it as close as possible functionally to the old software, but feel free to drop suggestions or requests in the Tech Support subforum!

Romney campaign admits it - They're in big trouble

The media will definitely sell the Romney comeback story to avoid a boring month. A few small time polls will serve that purpose.
 
And the over under on Democrats hitting the polls at the 2008 level of 39% to the Republicans 32%? That is the assumption of CNN's poll... and yet Romney and Obama are statistically tied given the margin of error.
 
PhDeac said:
The media will definitely sell the Romney comeback story to avoid a boring month. A few small time polls will serve that purpose.
Click to expand...

CNN is a small time poll? Rassmussen? Which poll was the most accurate in 2008 when all was said and done?
 
Pollsters don't select how many republicans and democrats are polled.
 
Fuck Your Couch said:
Pollsters don't select how many republicans and democrats are polled.
Click to expand...

Oh, really. Then why are they trying to model it on 2008 turnout? If you don't model the poll to reflect likely voter turnout it's really just propoganda. At least that is what Harrison Hickman would say.
 
DirkTheDeac said:
Oh, really. Then why are they trying to model it on 2008 turnout? If you don't model the poll to reflect likely voter turnout it's really just propoganda. At least that is what Harrison Hickman would say.
Click to expand...

Well Dirk, are they modeling it on likely turnout, or 2008 turnout? Have you studied their sampling methods?
 
DirkTheDeac said:
Oh, really. Then why are they trying to model it on 2008 turnout? If you don't model the poll to reflect likely voter turnout it's really just propoganda. At least that is what Harrison Hickman would say.
Click to expand...

Wow, you think that pollsters call people looking for democrats or republicans to hit a certain percentage when polling people? Wow.
 
I think he's talking about the weights used in the polls.
 
Good grief. I give up. Believe the polls. We'll see how that works out for you in four weeks. Better still, go to an SEIU meeting and do your own random sample of likely voters. Because FYC says pollsters can't control the sample or the methodology.
 
PhDeac said:
I think he's talking about the weights used in the polls.
Click to expand...

They don't weight the polls for party identification either. They look for X number of random responses. During the calls they ask for their party identification. Think about the political movements over the past 4 years and think about whether a conservative or liberal person is more likely to say they are independent.
 
Last edited:
DirkTheDeac said:
Good grief. I give up. Believe the polls. We'll see how that works out for you in four weeks. Better still, go to an SEIU meeting and do your own random sample of likely voters. Because FYC says pollsters can't control the sample or the methodology.
Click to expand...

It is against polling protocol to look for party identification or weight the polls according to party identification.
 
If you are pulling a true random sample how do you continuously sample 7-11% more Democrats when Gallup says the variance between party affiliation is roughly half that? What kind of random sample is that and would you care to estimate the odds of continuously finding a "random" sample that is always twice the estimated variance? Go for it.
 
DirkTheDeac said:
If you are pulling a true random sample how do you continuously sample 7-11% more Democrats when Gallup says the variance between party affiliation is roughly half that? What kind of random sample is that and would you care to estimate the odds of continuously finding a "random" sample that is always twice the estimated variance? Go for it.
Click to expand...

It is an attitudinal question, they aren't looking up party registrations and calling up 39% democrats and 31% republicans. They also aren't weighing it to assume 39% democrats and 31% republican or whatever. That would be freaking dumb and I can't believe people actually believe that's what the pollsters are doing.

"Umm, hey Bob what do you want to put the party identification at today?"

How can you not understand that right now more people will say they are democrats in a random sample than say they are republicans. That's because there have been notable conservative movements the past 4 years distancing themselves from the republican party. Those people will say independent when given the question "Do you identify yourself as a Democrat, Republican or independent?" BTW, there are more registered democrats than registered republicans as well.
 
Last edited:
Here is the Gallup data on party affiliation: Poll

I don't argue that it is impossible to poll a sample that is 11% more Democrat than Republican. What I do think is quite impossible is that multiple polls "oversample" Democrats consistently when you match their Democrat to Republican variance compared with the volatility of the Gallup data. Not to mention, I believe the most recent ABC poll included only 1% independents. Now why would that be? According to Gallup the majority polled consider themselves independents and according to CNN Romney currently is 8% ahead of Obama with independents. So, you can look at it in a number of ways, either the polls over the last month were engineered to shape voter mindset, or the methodology was unintentionally flawed, or the polls were pure guesswork and may mean nothing, or they may be dead on as we are led to believe.
 
DirkTheDeac said:
Here is the Gallup data on party affiliation: Poll

I don't argue that it is impossible to poll a sample that is 11% more Democrat than Republican. What I do think is quite impossible is that multiple polls "oversample" Democrats consistently when you match their Democrat to Republican variance compared with the volatility of the Gallup data. Not to mention, I believe the most recent ABC poll included only 1% independents. Now why would that be? According to Gallup the majority polled consider themselves independents and according to CNN Romney currently is 8% ahead of Obama with independents. So, you can look at it in a number of ways, either the polls over the last month were engineered to shape voter mindset, or the methodology was unintentionally flawed, or the polls were pure guesswork and may mean nothing, or they may be dead on as we are led to believe.
Click to expand...

Why are you so worked up about polls? I think you may spontaneously combust before November 7.
 
Question Dirk: For all those polls that generated that Gallup data about party identification, do you think something other than Gallup getting X number of responses, asking the party identification question and then recording that data happened?
 
I also find it funny that the poll doubters keep touting Romney's independent leads (amazing how the polls consistently get that right) and bitching about the party identification splits when Romney's independent leads helps explain the party identification splits.
 
I don't understand why national polls even matter one single bit. The only ones that matter are the ones in the swing states. Hey look at all these people in states that no matter what will vote one way or the other. The polls to look at are FL, OH, etc... thats it, nothing else. Obama can have the 70% latinos all in California, and Romney can be tied because of the 8 million Georgians voting for not the black guy, and it does not matter.
 
Skydog Deac said:
Why are you so worked up about polls? I think you may spontaneously combust before November 7.
Click to expand...

Just be glad the ACA forbids denial of service for is preexisting condition.
 
You must log in or register to reply here.
Back
Top