Call me odd, but I treat polls with a sample size of under 1000 people as little more then fun numbers. Pbrower's map is a fairly good indicator of why.
That is
literally the equivalent of saying "Call me odd, but I don't believe in the validity of calculus."
That sounds great but think about it. Road construction could conceivably block every single one of those people from going to the polls.
No, it conceivably could not. Such a thing would require a natural disaster whose scope is so terribly immense that the outcome of the 2012 presidential election is inconsequential.
Besides, even if the 1:100,000,000,000,000,000 chance of this did come up, it wouldn't matter -- these 406 people are a
representative sample of 3,000,000. The 2,999,594 voters NOT affected by road construction would go to the polls and vote exactly the way the poll predicts, plus or minus 4.9% in the 95% confidence interval.