Nate Silver debunks the widely held view that Rasmussen's GOP-leaning "house effect" is a result of its use of a likely voter screen earlier in the cycle than other polling outfits. He observes:
The bottom line is this: the sample included in Rasmussen's polling is increasingly out of balance with that observed by almost all other pollsters. This appears to create a substantial house effect, irrespective of whether Rasmussen subsequently applies a likely voter screen.
It also appears to be a relatively new facet of their polling. If one looks at the partisan identification among all adults in polls conducted in September-November 2008, Rasmussen gave the Democrats at 6.5-point edge, versus an average of 8.7 points for the other pollsters; their house effect was marginal if there was one at all.
It's interesting that the gap between Rasmussen's 2008 generic ballot polling and everybody else's was fairly small at the tail end of 2008, but it might not prove that the "house effect" we're seeing today is new. I haven't looked at generic ballot data, but I have looked through McCain-Obama polling and it appears that the gap between Rasmussen's polling and everybody else's narrows as election day approaches.
For example, from February to April of 2008, Rasmussen showed John McCain leading by an average of 2.6%. Meanwhile, every other poll of likely voters during that time period showed Barack Obama with a 2.6% lead. So from February to April, there was a 5.2% gap between Rasmussen and everybody else.
From September through election day, however, the gap was significantly smaller. The average Rasmussen poll taken during that time showed Obama leading by 3.8% while the average of all other polls of likely voters showed Obama leading by 5.8%, resulting in a gap of 2%.
So in the early months of 2008, the gap between Rasmussen and everybody else was 5.6% compared with 2% late in the cycle. Another important point here is that early in the cycle, Rasmussen's polling accounted for a far larger share of overall polling, meaning that their results played a more important role in shaping horse race narratives in the early stages of the campaign than they did in the latter stages of the campaign.
The shifting size of the "house effect" is certainly something worthy of additional exploration.