Quote:
Originally Posted by transplant99
|
That's one I haven't seen--seems to provide a decent summary of what's out there. I've been looking at Chris Scilizza's blog on Washington Post's website, and at the "election scorecard" on slate.com, which is pretty interesting, but uses averages of different polls--somewhat problematic given the different methodologies that are in use.
Polls are more art than science; a lot of pollsters actually don't use their raw numbers, but figure they get a more accurrate tally if they correct for voter party affiliation. What that means is that if 62% of voters in 2004 were Republicans, they'll actually use 62% Republicans in their final sample, regardless of what the raw data are. A weird practice, IMO--but there you have it.
What it means is that voter turnout is a huge X-factor. Which party gets their base out more is going to determine whether this turns into a 40-seat wave or a 10-seat ripple. Historically, Democrats have been bad at this for the last few election cycles. Hopefully (if you're rooting for them) they've figured it out.
One point of interest: conventional polling methodology says that in close races where the incumbent polls lower than 50% in the final weekend, the incumbent is in trouble, because late deciders tend to break for the challenger. If that's so, look for Democratic pickups in Montana and Virginia. I also think they'll eke one out in Missouri (my neighbourinos to the South!)--so it will come down to winning one of Tennessee or Arizona, if my math is right--but both of those are long shots. Corker (R) is polling higher than 50% in Tennessee, and though the GOP candidate is polling lower than 50 in Arizona, he does have about a 7-point lead, I think.