The polls were dead on, or at least the good ones (e.g Rasmussen , Battleground and IBD/TIPP). Gallup, Pew and CBS/New York Times pollsters can now leave the room. McCain lost by 6%, after leading by 3% in mid September. The national numbers moved 9% after the financial crisis hit, and so did the battleground state numbers (a bit more or less depending on the state).
What states does McCain win if his numbers improve by 9% net based on the results yesterday: Florida, Ohio, Virginia, North Carolina, and Indiana easily, and Colorado close. That gets McCain to 269-269. A tie in this case would have gone to the winner last night -- Obama, due to the Democrat's edge in the House of Representatives. McCain in the end would have needed to win all these states and either Nevada (a surprising blowout last night for Obama) or Iowa (closer than expected). Would we have felt better today to win the popular vote, tie in the Electoral College, and lose in the US House? I doubt it.
We can always wonder what would have happened last night if the financial crisis had hit in mid November, and not mid-September, and of course, why it happened when it did.
At the moment, McCain is losing by 7 million votes: 80% of that gap comes from Obama's margin in just 3 states: California, New York and Illinois. In 2004, Bush won by 3 million votes, so the net shift in 4 years was 10 million votes. As best as I can estimate today from the exit polls and actual results, the 10 million vote shift was primarily among two groups: African Americans (net 5 million additional vote margin for the Democrats -- 13% share of the vote, and a 95-4 margin for Obama), and Hispanics (net 3 million additional vote margin for the Democrats -- 9% share, and a 68-31 margin for Obama).
Total turnout for President came in nowhere near the record numbers some were forecasting: 140 million or more. The total vote for President might not exceed the 2004 level of 122 million by very much at all (currently 118 million with 96% of the votes counted). What that means is that with higher turnout by African Americans and Hispanics, about 2% higher as a share of the total vote in each case, that white turnout was down from 2004. Some Republicans (and Democrats and independents) stayed home. There is no other conclusion.
Republicans appear to have gotten lucky in the Senate races: Saxby Chambliss appears to have won in Georgia with just over 50%,avoiding a runoff. Norm Coleman is leading by fewer than 1,000 votes in Minnesota against the execrable Al Franken (there must be a God, if this holds). Gordon Smith is hanging on in Oregon by about 15,000 votes, though a quarter of the all mail vote in that state has not yet been counted. My quick survey of the counties not yet fully in, suggests Smith will hold on. And then there is Alaska, where Ted Stevens may have lived to fight another day, both in court appeals and then to hold his seat in the Senate from attempts by other Senators to force him to resign, assuming his current lead holds up after absentee votes are counted.
In the House, Democrats did well, but not quite as well as expected, picking up a net 22 seats.