The Fox national poll showing Mitt Romney up by one point is significant not because it means Romney is “ahead,” but because it reminds us that who you sample and when you sample make all the difference in the world. Was Romney really behind by nine points earlier in the month as liberals crowed, and did he really go up 10 points in two weeks? Well, maybe the nine-point poll wasn’t really accurate, and then again Fox went from registered voters to likely voters. (Among registered voters in the most recent poll Romney is down by two points.) This is why agonizing or celebrating over one or another poll (especially national polls) is rather silly.
Each side will be spinning furiously with each batch of poll results. A tiny bump for the other guy! Look, no bump after the first debate! Rather than squint at each poll, trying to discern its accuracy and decipher the accompanying spin, I’d suggest you look at a few factors.
First, look at President Obama’s approval rating; which candidate wins on handling the economy; and right track/wrong track. Those factors alone are probably more reliable than the horse race numbers.
Second, is the electoral map expanding or shrinking for Romney? Right now, it seems to be expanding, showing him within the margin of error in Iowa, Nevada, New Hampshire and Wisconsin, for example. Remember when pundits were saying Romney had a “narrow path” to 270 electoral votes? Now he has a bunch of paths (Ohio’s 18 electoral votes can be more than made up by Wisconsin’s 10, Nevada’s six and New Hampshire’s four electoral votes.)
Third, national polls showing a “big lead” for Obama among Hispanics or other key constituencies are virtually meaningless. To the extent that they are measuring Hispanics or other core Democratic constituencies in deep-blue states like New York, California and New Jersey, they won’t tell us anything useful. (These states are already going to Obama.) It is the Hispanic vote in swing states that is key. Moreover, Romney need not win the Hispanic vote, or young voters, or Jewish voters; but if he does better than Sen. John McCain (R-Ariz.) did in 2008, that is significant.
Finally, every pollster and political consultant I have spoken with tells me they expect the electorate to look more like 2004 and 2010 (the electorate has roughly equal numbers of Democrats and Republican) than in 2008, when Democrats enjoyed a seven-point advantage. Polls with closer splits between Republicans and Democrats are likely more predictive.
There is a big caveat that goes with any and all polls. The news cycle is so fast, and the events of the campaign come so quickly (the two conventions are back to back) that by the time you poll and report results, the race may have taken a new turn. In a campaign moving at 100 mph, all the polls give you is a blurry snapshot out the back window.