Living In Rasmussen-Land

Andrew Sullivan —  Aug 4 2012 @ 2:52pm

If Drudge is flagging a poll, you know almost instantly who has provided it. But no one has more completely swallowed the Rasmussen world's alternate reality than John Hinderaker. He acknowledges that in the Rasmussen poll, there are more Republicans in the sample than in most others. But he runs with it anyway. Romney is clearly winning the race, it seems.

The Rasmussen effect is best seen by using various polling models and removing Rasmussen data to see what happens. So's data, including Rasmussen, currently shows a national tie at 46.1 percent. Removing Rasmussen makes it Obama 47 and Romney 45.7. Small – but in a race this tight, not trivial. I generally remove Rasmussen from the poll of polls, because they are so openly biased in their sample. But whenever I cite a specific poll, I try and place it in the broader context of all the others.

This would be a problematic strategy for Hinderaker. Rasmussen nationally gives Romney a 4-point lead. Gallup – which consistently pegs Obama lower and Romney higher than most – gives Obama a 2 point lead. RCP gives Obama a lead of three points. The poll of polls without Rasmussen gives Obama a 1.3 percent lead. The difference – 5.3 percent to 7 percent – is huge. Hinderaker writes that according to Rasmussen:

Romney has generally led Obama ever since he became the clear front-runner for the GOP nomination.

Gallup shows a pretty even race until late June, after which Obama has a small but steady lead. RCP's poll of polls shows that Obama has never lost a national lead. But when you're busy creating an alternate reality, you don't want other data to interfere, do you?