Politics: Rasmussen and Gallup Diverge On 2012

Rasmussen and Gallup Diverge On 2012 ... at the moment, as of this Friday morning, as I look at the RCP numbers. Gallup is saying Romney by 7% and Rasmussen has it tied. Rasmussen claims an MoE of 3%, so the tie could really be a 6 point lead for either side. Gallup cites an MoE of 2%, so Romney's lead could be as big as 11 and as small as 3.

Why are the two polls so divergent?

The likely voter screens they apply are different. But I wonder what aspect of their screens is leading to such different results? Any polling experts out there who can offer a hypothesis as to why the 2 big name polls are so different at this moment?

We shall see if the divergence continues.

At the moment, on RCP, Gallup is the "outlier" data point with the race being much closer if one aggregates the polls.

UPDATE: Jay Cost of the Weekly Standard was on the Hugh Hewitt radio show and suggested that Gallup's use of a tight likely voter screen ironically makes it more volatile to momentum shifts. Thus, the direction of many polls in favor of Romney results in Gallup perhaps overstating the movement. Anyway, that was the impression I got from what he was saying in the interview.