Rasmussen sees a different world

First, the composite Party ID of all pollsters:

Independent: 41.5%
Republicans: 25.8%
Democrat: 30.3%

Next, the view from Rasmussen:

Independent: 30.4%
Republicans: 35.5%
Democrat: 34.2%

Now, all pollsters except Rasmussen:

Independent: 41.4%
Republicans: 25.9%
Democrat: 30.2%

So we notice that Rasmussen sees a much more Republican country out there than other pollsters. And Rasmussen can't claim that his mysterious "likely voter" screen is making the difference, since Rasmussen's party ID figures are raw (all adults) figures. All of the figures shown for other pollsters are also an "all adult" sample.

Rasmussen is surveying a country with nearly 10% more Republicans, 10% fewer Independents, and only 4% more Democrats than other pollsters.

PPP surveys registered voters. Let's look at what their figures indicate:

Independent: 25.8%
Republicans: 35.7%
Democrat: 38.2%

All pollsters, excluding PPP, shows this:

Independent: 29.6%
Republicans: 31.2%
Democrat: 37%

This is much closer to the composite than Rasmussen was with all adults. About 4% fewer Independents for PPP, and about 4% more Republicans than the others are finding. The numbers for Democrats is about the same. So, in spite of being labeled as a "Democratic" pollster, PPP is looking at more Republicans than their peers polling registered voters.

The most current poll from GWU shows "likely" voters to break down this way:

Independent: 27%
Republicans: 34%
Democrat: 37%

In my opinion, it's still too early for any "likely voter" (LV) screen to be worthwhile. But Rasmussen's claim that his LV screen is the reason for his outlier status is not borne out by looking at another pollster that is also using an LV screen. Apparently Rasmussen just happens to consistently hit on a gold mine of unhappy Democrats in their sampling, judging by their results.

Since Rasmussen is working from a much more Republican-based sample to begin with, it's hardly surprising that they are skewed toward "unhappy Democrats" in their LV sample. We don't really know what Rasmussen means by a "nationwide" sample. My guess is that Rasmussen's "nationwide" sample includes a lot more rural areas and lot more "red" states than other pollsters.

Rasmussen's 2011 Christmas polling (12/27-12/28) showed Romney with an outlier 6 point lead for Romney over Obama (45/39). Nobody else is seeing that.

As we've seen multiple Republican candidates "flare up" with popularity, Rasmussen has always mysteriously found a 2-point edge over Obama for each person at their peak. Rasmussen said Bachmann, Cain, Perry, etc. would all beat Obama in a match-up by 2 points. Later, the numbers change as the candidate slumps. 2% is such a weasel figure: within the margin of error, and not so large that it could seem suspicious if it switched to an Obama lead.

We're at a point in the cycle where Rasmussen can publish outrageous results, and then later say that "things changed". This is what usually happens with Rasmussen polling, whether it's a national race or a Senate contest. The Republican always shows a good lead at this point, and that lead always seems to diminish or flip as we approach the actual election. Rasmussen failed to do this "reversion to the mean" in 2010, and got stuck with an 8-point Republican bias in polls close to the election. With that, pollsters are much less intimidated to publish figures contradicting Rasmussen than they used to be.

It should also be pointed out that PPP publishes the raw figures (including area codes) for all of its polling. PPP also asks each person surveyed to place themselves in one of four regions: South, Midwest, East, or West. They publish the breakdowns of the sample in every poll, so you can see if they are oversampling one region. By examining the area codes, you can see if PPP is oversampling rural areas or red districts. Both PPP and Rasmussen use automated polling and call landlines. Yet, Rasmussen sees a very conservative America where even Democrats agree that Obama is the source of all our troubles. Oddly, Rasmussen is anything but transparent about his figures.

PPP has done an immense amount of polling for free. They have established a huge amount of base-line data that will be extremely valuable going into the 2012 election. No one paid for these polls; PPP volunteered. Anyone discounting these polls due to PPP being a "Democratic" polling firm is denying themselves the facts. There is also a lot of other polling that PPP has done where it's been paid for by DailyKos/SEIU. Critics love to mix these two groups of polls together to claim a bias and taint the free polling's validity. Don't let them confuse you. Aside from that, Rasmussen makes his money from subscribers that are overwhelmingly right-wing. He also does speaking gigs where he pushes right-wing narratives, so he has a strong incentive to produce dramatically pro-Republican numbers. I'm not saying Rasmussen is dishonest. I'm saying that if you dismiss PPP based on financial bias, you would have to dismiss Rasmussen as well.

By the way, let's look at Gallup's party identification numbers.

Independent: 40%
Republicans: 27%
Democrat: 31%

These figures are an "all adults" sample, like Rasmussen. Gallup also calls cell phones in their sample and uses live interviewers (not automated polling).

There was a two-point increase in independent identification from 2010 (38%) to 2011 (40%). The increase in independent identification came at the expense of Republican identification, which dropped from 29% to 27%, while Democratic identification held steady at 31%.

Gallup also says that, if you include "leaners", that the country splits 45/45 on D/R. That leaves only 10% real "Independents" to determine the election. The scary part is that this is pretty much how things broke down in 2010. The difference is in registered voters, of course. It's ironic that the Republicans screamed about Obama "not listening" to the majority, when things were actually evenly divided among "The People".