This little chart shows that presidential polling in 2004 was the most accurate of the last quarter century.
In fact polling has since 1980 gotten progressively more accurate each election with the exception of 1996, when polls famously over-predicted Clinton's margin of victory over Dole. This will lead some to wonder whether the same thing is happening now. Perhaps. But I doubt it. Clinton was the incumbent president, which makes for a completely different psychology. I think.
My sense is that things will be a little tighter than the polls suggest, but only a little -- two or three points nationwide, meaning a 4.5% to 5.5% Obama win rather than the current 7.5-ish. This would prevent Obama from winning some states but shouldn't interfere with the quest for 270 unless things get really weird out there.
On the other hand if this Obama ground game really blows the gates off things, then I guess he could win North Carolina, Missouri etc and turn this into a blowout. That's one of the things I'm most intent on watching for today and tonight. Every anecdotal thing you read -- from Kos diarists, say -- sounds pretty formidable. We'll see.
.jpg?w=600)
