Nate kicked off last night's poll round-up by observing that, "Although the national tracking polls are trending upward for Obama, this set of state polling is a strong one for John McCain."

Relatedly, last week, while Obama's national numbers seemed to be cratering, a colleague talked to a Democratic operative with access to state-level internal polls, who said the state numbers were still looking solid for Obama.

We've seen this a couple times this summer, which leads me to wonder: Why do state polls (particularly swing-state polls) seem to lag behind national numbers? After all, you'd think swing-state voters would be paying closer attention to the campaign, and therefore processing information more quickly.  

I can think of a couple of theories:

1.) Because there's much more political advertising directed at swing states, actual news events and the changing media CW take a little longer to break through there. Put simply, swing state voters are surrounded by a denser fog of war than the rest of us.

2.) Undecided voters tend to live in swing states, and they follow the news less closely than partisans, who live and die by every twist and turn of the campaign. Problem is--do we know swing states have more undecided voters? It sounds logical, but a state could be a swing state because 49 percent of voters are dug in on one side and 49 percent on the other, leaving 2 percent up for grabs. On the other hand, a safe state could be 70 percent pro-McCain or Obama, 15 percent for the other guy, and have 15 percent undecided. It's really tough to say (and the polls themselves aren't always helpful, because some pollsters push undecided voters harder than others).

3.) The methodologies for state and national polls frequently differ. The national polling we consume on a daily basis are generally tracking polls, while the state polls tend to be more traditional polls, which take a little longer to turn around. So maybe the national polls are just fresher. (The downside, I guess, is that they may be overly sensitive to short-term developments.)

Anyone have any other ideas? I'd be curious to hear what Nate and Mark Blumenthal think. Apologies if they've written on this recently and I somehow missed it...

--Noam Scheiber