Skip to comments.GOP Increase in Party ID After Convention Not Unusual
Posted on 09/11/2008 5:50:59 AM PDT by bcatwilly
For Freeper poll watchers, this is an interesting article about party identification. This is really the reason that Rasmussen has not shown the same gap for McCain that Gallup has recently, as Rasmussen is weighting his poll according to a full 90 days prior survey of what he believes the makeup by party will be come election day. I guess that we will all find out on November 4th, but I personally don't think that this is a traditional situation where the increase in Republican party identification or leaning will decrease to levels that Rasmussen is currently weighting his polls to. I tend to lean more toward the fact that Governor Palin just woke up a few people that wanted a reason to identify more with that side, myself included. It will be a tough battle, no question of that.
(Excerpt) Read more at gallup.com ...
My poll says that America will not elect a black, Marxist-socialist president...
Never saw these stats before, thanks.
Before the GOP convention, 39% of Americans said they identified with or leaned to the Republican Party, but that number has increased to 47%. Forty-eight percent now identify with or lean to the Democratic Party, down from 53% prior to the GOP convention.
Never had trusted Ras. He had Romney leading for the longest time. Even when we all knew it was not going to be Romney.
He’s got as much special sauce in his system as Zogby.
The polls are used to sway public opinion and nothing more
Remember the so called ‘snap’ polls? Where are they now?
Surely someone snap polled Sarah’s convention speech!
that is why some polls that use a 90 day rolling average for party identity and weight their polls pick up movement very very slow. Gallup when doing their truly random sample picks up where the people are today.
Maybe some that know better than I can answer this, but I thought that standard polls that weight by party historically have used the party identification numbers as reported in the actual exit polls from prior elections.
I thought that the 90 day rolling party ID approach was somewhat unique to Rasmussen, but I could be wrong. I do think that there is some truth to increased Democratic party identification, but the Palin factor of energizing the conservative base cannot be diminished either.
One thing that I found mildly interesting at least was a question asked in the latest FoxNews/Opinion Dynamics poll of those who have registered to vote since 2006, and Obama only had a 2% edge among that group of voters. Now I guess that a poll can’t account for all of the “new” Democratic voters who are either currently 6 feet under or roaming the streets waiting to be offered a pack of smokes on election day to hop a bus to the polling station :)
Sure. Let me clear this up. Gallup and SurveyUSA DO NOT use weighted sampling. Zogby and Rasmussen do. An example is this: If you have 100,000 black and white jelly beans in a jar and you want to determine breakdown you don’t need to count them all..You need to mix them up fairly (good random sample) and pick 1000 of them. Based on this (assuming picked fairly) then you will get a good indication of the breakdown of the 100,000 jelly beans. Picking 1000 of them versus 500 of them give you a smaller margin of error. In contrast, weighting means you pick the 1000 jelly beans but you assume 39 percent are black, 29 percent are white, and 32 percent can be either (this is weighted). Problem with weighting is that party identification in polls changes they are dynamic and not a true random sample. This is why CBS polls etc...are so off so many times. They weight versus random samples. SurveyUSA and Gallup are truly better approaches to polling in my opinion. I keep this opinion whether poll is good or bad. I am a statistics major and understand this subject very very wel..
The Gallup numbers suggest a seismic change in the electorate following the GOP convention. While some of that may be a temporary bounce that will fade, if the number of people identifying with each party stays close to even, we will win this election handily.
Thanks for posting this. Very encouraging news.
Yes, 48% to 47% ID was shocking.
Maybe this info really needs its own thread.
I would not agree, but only because I look around FR and see dozens of conservatives who a month ago said they wanted nothing more to do with the Republican party, who now are happy to say they are McCain/Palin supporters, and I’m guessing are now calling themselves Republicans again.
Which means if you polled them a month ago, they would have said they were “independents”, and if you polled them today they would say they were “republicans”.
If so, Rasmussen would throw them out of the poll now because he would see he polled too many republicans based on his model.
It would be interesting to see what Rasmussen’s daily “discard” rate of republicans are right now, to see if in fact he has had to correct more and more in the past two weeks for an “oversampling of republicans”.
I don’t think this is a temporary phenomena. A lot of people decided they didn’t like being called republican after the 2006 elections. But with the Drill Here movement, the unpopularity of the democratic congress, and McCain’s takeover of the party, I think a lot of people are more than happy to be called republicans now.
As I just attended a rally of over 23,000 people on a wednesday morning who stood for hours to stand at a republican event and cheer the republican candidate, I think it’s reasonable to conclude that there are more people today who would tell a pollster they are republican than there were two or three months ago.