Perhaps someone can explain how Gallup is more accurate? I see four interesting things.
1. Gallup's numbers add up to 93%; Rasmussen is 100%.
2. USA adds up to 94%.
3. Does meeting an average make it MORE correct?
4. Gallup and Rasmussen are more current polls.
I interpret these numbers to mean Gallup and the others allow an undecided category, where Rasmusen requires an approve or disapprove response. The differing methodologies are going to cause some level of distortion. It will be interesting to see of Rasmussen's is a predictor of the direction we will go when people make up their minds.
One big methodological issue is that Gallup tracks approval/disapproval without any division inside those groups. Rasmussen's reported approve/disapproval comes from summing the figures for "strongly approve" and "somewhat approve" subsets. The result of adding those options is a rush to moderation. If someone is unsure, but feels strong enough to have a response, they may consider themselves as someone who somewhat disapproves, but if asking approve/disapprove, they'd say approve. Works the same way for low approval ratings too, but in the opposite direction (somewhat approve/disapprove).