the polls, as a whole, were RIGHT ON THE MARK.
They were incredibly accurate in many instances--the VA race being a good example--as the RCP poll aggregation showed Webb with a tiny lead on Allen--exactly the outcome that occurred.
The vast majority of the polls told us exactly what we would see on election day--and only a small percentage ended up being outside the margin of error. Regardless of the biases (real and perceived), most polls were CORRECT.
I think every one should remember that next time--when everyone starts trashing polls because they don't like what they say.
I was surprised by this too. In MI, polls have been famously OFF for gubinatorial campaigns. Yet this time they showed a 10+ pt spread for Granholm, pretty accurate prediction. Have the pollers actually improved methodologies, or is this a case of the "stopped clock right twice a day" syndrome?
Even pundits who always make wrong predictions (Bill Kristol, Dick Morris) got it right this time.
Just a few thousand votes the other way would have made the polls totally wrong.
A large number of independents (it was said) did not even decide on who to vote for until a couple of days before the election, and most broke for the Democrats.
When the GOP wins lets see how accurate the polls are on that one.