It's increasingly hard to take Rasmussen seriously...

[Drunken Irishman]

His current head-to-head tracking poll has Romney beating Obama by 5. Just a couple days ago, Obama was up 8.

Am I to believe that, in a span of maybe three to four days, Romney improved his overall numbers by 13 points? That seems ridiculously high. Remarkably high. For numbers like that, you'd expect that Obama murdered someone on live television. It just really doesn't happen - not that quickly, anyway.

What will Romney's numbers be tomorrow? 10? Tied? They're so erratic, it's hard ot take those numbers as serious. This race is fluid and because it's fluid, a jump of support like that makes absolutely no sense.

[NJMaverick]

I wonder just how big the margin of error is

[Peacetrain]

Now if I remember correctly DI, I think this is part and parcel of Rasmussen.. He will put out the outlier poll.. (which many of the right cling to like a child's blanket) and then a week maybe 5 days before an election.. change his poll up to basically match everyone elses polling to try and keep some kind of relevancy .

[jaxx]

(03-11-2012 09:43 AM)Peacetrain Wrote:  Now if I remember correctly DI, I think this is part and parcel of Rasmussen.. He will put out the outlier poll.. (which many of the right cling to like a child's blanket) and then a week maybe 5 days before an election.. change his poll up to basically match everyone elses polling to try and keep some kind of relevancy .

I read an article about this last night....should have posted it because I can't remember the two ways he polls. But they are at odds with each other and he is switching to the other one now.

Rasmussen was the lagger in 2008 and I look for him to be this time too. And Gallup is just as crazy this year.

[Peacetrain]

(03-11-2012 10:55 AM)jaxx Wrote:  

(03-11-2012 09:43 AM)Peacetrain Wrote:  Now if I remember correctly DI, I think this is part and parcel of Rasmussen.. He will put out the outlier poll.. (which many of the right cling to like a child's blanket) and then a week maybe 5 days before an election.. change his poll up to basically match everyone elses polling to try and keep some kind of relevancy .

I read an article about this last night....should have posted it because I can't remember the two ways he polls. But they are at odds with each other and he is switching to the other one now.

Rasmussen was the lagger in 2008 and I look for him to be this time too. And Gallup is just as crazy this year.

I would like to read that jaxx if you find it.. I just remember in 2008 Rasmussen switched it up right at the end.. so now it is because he basically is keeping 2 sets of books as it were.. makes you want to go hmmmmmmmmm

[jaxx]

(03-11-2012 03:01 PM)Peacetrain Wrote:  

(03-11-2012 10:55 AM)jaxx Wrote:  I read an article about this last night....should have posted it because I can't remember the two ways he polls. But they are at odds with each other and he is switching to the other one now.

Rasmussen was the lagger in 2008 and I look for him to be this time too. And Gallup is just as crazy this year.

I would like to read that jaxx if you find it.. I just remember in 2008 Rasmussen switched it up right at the end.. so now it is because he basically is keeping 2 sets of books as it were.. makes you want to go hmmmmmmmmm

This isn't the article, but give you an idea of how he polls.

<..> Just a couple problems: As it happens, Rasmussen is a former Republican pollster. The company also uses weird logic to conduct its polls.

Rasmussen uses an “approval index” to figure out the president’s numbers, rather than asking whether or not those being polled just approve of the president. For instance, the pollster’s latest, conducted yesterday, shows 26 percent of voters “strongly approve” of the president, while 39 percent “strongly disapprove.” Therefore, the index is -13. Make sense?

Santorum is not surging nationally, though polls in this case show that he—and only he—leads Obama in a head-to-head matchup, amongst all the presidential nominees. Rasmussen polls regularly show conservatives ahead in close races, or at least close in not close races. Whether or not that has anything to do with the strongly-approval index is hard to tell, but Rasmussen only includes “voters” in its surveys—and one is much more likely to call him/herself a Republican voter than a Democrat one at this stage in the game. The polls are regularly accused of having a conservative bias.

http://blogs.philadelphiaweekly.com/phil...ustworthy/

[There Is No Spoon]

Lots of ways to skew a national poll. Skewing the individual states that are required to win isn't such an easy feat, though I'm sure are friend at Republisson is doing his best there as well to keep Republicans from despairing.

[Willinois]

Does anyone take the Koch Brother's favorite pollster seriously anymore?

[Drunken Irishman]

A lot do, unfortunately.

[Drunken Irishman]

You can view Rasmussen's poll numbers from the '08 race here. Unfortunately, the numbers only go to October 15th, as that was the latest date I could get on the election.

You'll notice, though, that in October, the entire month, actually, Obama led McCain. His highest lead was 8 on the 6th and 7th - while his narrowest lead was six on the 1st. He was leading, though, for the last part of September. In fact, the last day McCain led Obama was 9/17.

Now RCP's average of the polls in '08 shows that Rasmussen had Obama winning the election 52-46, or +6.

But those polls don't appear to be released every single day. So, who knows in that regard.

Anyway, let's compare the Rasmussen numbers with RCP's average from August 1st to October 1st. I think that'll give us an indication of how similar or off the numbers were. And I'm gonna do every few days so that it's not too terribly long.

(R: Rasmussen, A: Average)

08/01/08 - R: +1, A: +2.6
08/05/08 - R: -1, A: +2.5
08/10/08 - R: +2, A: +4.4
08/14/08 - R: +1, A: +4.2
08/18/08 - R: +1, A: +3.6
08/20/08 - R: +1, A: +1.2
08/24/08 - R: +3, A: +1.7
08/28/08 - R: 0, A: +2.9
08/30/08 - R: +4, A: +3.9
09/01/08 - R: +3, A: +5.5
09/05/08 - R: +2, A: +2.6
09/07/08 - R: 0, A: -1
09/10/08 - R: +1, A: -2.2
09/14/08 - R: -3, A: -2.1
09/17/08 - R: -1, A: 0
09/20/08 - R: +1, A: +2.3
09/24/08 - R: +2, A: +3.5
09/28/09 - R: +6, A: +4.8
09/30/09 - R: +6, A: +4.8

Looking at these numbers, the difference isn't huge. In fact, the average of Obama's lead in the polls in Rasmussen through this stretch was 1.5 and the average in the average was 2.3 - or .8% - less than an entire percentage point.

With that in mind, it's hard to say Rasmussen's numbers lagged that much. Then again, .8 in a relatively tight race, as it was back at that point in '08, makes a big difference - especially when it's possible those same Rasmussen numbers were being weighed in the daily averages.

Plus, I didn't look at it every day, so it's certainly not exact. But you get the idea.

My problem with Rasmussen isn't that he's showing Obama trailing. It's that now Romney has done a complete 180 in support since Obama was leading by roughly eight points last week. For something like that to happen, you would think there would need to be a massive scandal because a double-digit swing in only a few days is almost unheard of, especially when there really is nothing going on in the news to force that change.

[SeattleGirl]

(03-11-2012 04:00 PM)Drunken Irishman Wrote:  My problem with Rasmussen isn't that he's showing Obama trailing. It's that now Romney has done a complete 180 in support since Obama was leading by roughly eight points last week. For something like that to happen, you would think there would need to be a massive scandal because a double-digit swing in only a few days is almost unheard of, especially when there really is nothing going on in the news to force that change.

I think that's what "Lies, damned lies, and statistics" means.....

[Drunken Irishman]

So, Romney's lead has shrunk to 3 points. See what I mean when I say it's erratic?