Re: Death to the Incumbent! Part Two: Now with more Death.
geezer and wewantmore,
The problem with both nate silver's blog and realclearpolitics is that they're relying on the frequent robo-callers who's accuracy won't be known for a few weeks. In this environoment, especially after what happened in Alaska, who the hell knows anymore? My general take on the Senate is as follows: In almost all elections, races tighten whether its Dems or GOP'ers leading as people come off the fence closer to election day. This affects both sides, but usually one candidate's supporters are more enthusiastic early but then some of the opponents supporters eventually come on board. This happened in both 2008 and 1996 where hapless GOP candidates closed somewhat at the end, as well as 2000 where Bush II's sustained lead evaporated the weekend before the election.
What makes this year different is that virtually all the people on the sidelines right now are Dem leaners. The GOP side is maxed out on enthusiasm. So they key to a big win for the GOP is keeping the unenthused home. That will happen to some degree, but as we get closer to election day more of these people will become "likely" voters. Why does this matter? Because in Senate races such as CA, WA, WV, CO, NV, and IL - so called "toss up" races having say half of the now non voting public re-engage, means its more likely those voters are Dems coming home and that takes these off the board. Somebody like Fiorina for example really needs Dem voters to think the election is on Wednesday instead of Tuesday.
So, where does that leave us race by race? ND, IN, AR are gone, although I don't think Lincoln will lose by the 20 all poll have here down by, losing by 10 vs losing by 20 = the same result, you lost.
WI & PA - poor campaigns by Dems put GOP in front in "blue" states. With Dems most likely down by mid single digits, they really need to push Dem engagement because the normal tightening of races towards the end probably isn't enough to get them there.
KY, AK, FL - wild card races. I don't trust polling in any of these. Way too many variables. Meek's support is going to fall off a cliff, the question is whether those people stay home or vote for Crist. Really, who knows about Alaska? For Kentucky, I'm curious whether out and out proud libertarian Rand Paul, a guy who I think unlike a lot of people would be true to his beliefs once elected, attracts the support polls claim he is in hugely government dependent Kentucky (similar issue in Alaska, but that state can choose another GOP'er whereas there's only the Dem alternative in Ky).