I mean, it is mathematically possible. But it's just not very feasible. I still have six or seven races that could certainly go either way, and all but MD are GOP seats. (Sorry, I just don't like the GOP's chances in NJ.) But in all those races, the GOP incumbent has a significant fundraising advantage (excluding Santorum - even, and Chafee - opponent has over twice as much). And Steele has about a half million more than Cardin.
Who can shed some light on this question? How do the pollsters adjust for the fact that Republican likely voters are more likely to actually vote than Democrat likely voters?
Zogby uses his "special sauce" but what do the others use?