Were The Polls And Exit Polls Accurate?

We don't know nearly enough yet to be sure.  But on balance, it seems like the polling industry should be happy with its work this cycle.

Most national polls got Obama's number right -- 52-53% -- but some  understated Sen. McCain's total by about three points (NBC/WSJ, CBS/NYT, Fox).  Pew and CNN called it almost exactly. So did, t must be said, the Rasmussen IVR, called it exactly.

Mason-DIxon's final state polls seemed to be right on, although Pennsylvania's margins did NOT narrow to the degree that almost every poll showed they had.  Nate Silver nailed it.

The unweighted exit polls as I saw them last night weren't bad.  In some states (VA and NC), the support for the Democratic candidate was overstated slightly,  but there were no egregious errors and the methods put in place by the consortium to reduce various biases seem to have worked.

I'd like to hear more from polling experts on why, in national surveys, major pollsters can't seem to figure out how to accurately survey the right side of the political world in these things as accurately as they do the left size. Non-response bias? Distrust of the media?