Mystery Pollster has an excellent piece up on the Rasmussen poll. It is lengthy, and I can't possibly summarize the entire piece here, so poll junkies should go check it out for themselves. I think this post goes a long way toward explaining the age-old question (age-old in blogosphere time, anyway): why does Rasmussen end up with results that lean to the right? Basically, the answer can be found in the significantly lower number of independents Rasmussen includes in their surveys compared to other polling firms.
MP writes:
The biggest difference is that Rasmussen has fewer respondents in the independent and other categories (29%) than most of the other polls of adults (an average of 39%). However, Rasmussen's independent/other result is very close to average of recent surveys that screened for registered voters or likely voters (LV - an average of 30%).
Now, we all know here at MyDD that
independents have long since soured on Republicans and on Bush. So, it should seem pretty obvious to everyone here that when a poll reduces the number of independents in a poll, that poll will trend favorably for Republicans, even thought he poll still gives Democrats an edge in overall partisan self-identification.
One word of caution: it might be entirely reasonable to reduce the number of independents in a poll and instead focus on a more partisan universe of voters. This is especially the case for a low-turnout, mid-term election. Combined, Democrats and Republicans will probably make up at least 72% of the vote in November, and might make up more than 75%. Compare that to recent polls, where they are only combining for an average of 61% of the estimated electorate, and you might see our problem.
Independents turn out at low rates, so the Democratic lead among independents might not be as important as many are currently construing it to be.
A more partisan universe improves Republican odds greatly in 2006. While
the last twenty-seven polls over at Polling Report show Bush with an average approval of 37.6% positive and 57.3% negative,
the last sixteen Rasmussen polls show an average of 42.4% positive, 55.9% negative. In terms of net approval, that is a 6.2% improvement for Bush in the more partisan Rasmussen universe. This is the type of swing that moves the election away from a landslide-realignment, and toward a close battle for control.
Since there are many good reasons to believe that the electorate will be more partisan than the general public, or even more partisan than the universe of registered voters, do not discount the Rasmussen results lightly.
2006 will be a base turnout election, and our party leadership needs to realize that
there are more to elections than independents. Emmanuel and Schumer seem to believe that
anger alone will be enough to drive up Democratic turnout. I am not convinced. We must
appeal to the activist base to build our political machine in ways other than simply watching Republicans implode.
