That’s the cautious analysis from one of the best in the business, Nate Silver of FiveThirtyEight at the New York Times.  Nate’s cautious because he rightly notes that we don’t have a plethora of data yet from swing-state polls, nor a real consistency in how they’re conducted in a broad sense.  However, Mitt Romney had been trailing in poll “wins” earlier this summer by a significant margin, and now has managed to close the gap almost entirely:

One simple thing I’ve noticed is this: Mitt Romney has held the lead recently in quite a few swing state polls. That wasn’t the case in June or July, when Barack Obama held leads in those polls far more often than Mr. Romney did.

Right now, our forecast model classifies the 10 most important states in the Electoral College as follows, and in the following order: Ohio, Virginia, Wisconsin, Colorado, Pennsylvania, Florida, Iowa, Nevada, New Hampshire and North Carolina.

If you simply count the scientific polls that have been conducted in those states since Aug. 1, Mr. Obama has led in 13 of them, Mr. Romney has led in 11, and there have been two exact ties.

This marks a contrast from June and July, when Mr. Obama held leads in polls of the top 10 states about four times as often as Mr. Romney did.

Keep reading, though, because the news isn’t all good.  The margins tend to favor Obama more, for one thing, and Nate questions the “Republican leaning” of some pollsters.  However, that would have been equally true in June and July, which is why the longer time frame matters and a simpler approach might tell a more accurate story.

If this is the case, though, then why have the national polls stalled at the mid-summer status quo?  That might be a function of Obama’s strength increasing in states he’ll win anyway, like New York.  The latest from Siena (a well-respected pollster) shows Obama up 29 points.  A month earlier, Quinnipiac had Obama up 23 points.  That’s bad news for Republicans in down-ticket races, but hardly a surprise.  It matters little to Romney whether he loses by five points or fifty in states like New York and California, except to the extent that Team Obama has to spend resources in those states to keep them in line.  If the deep-blue states get deeper blue, it would tend to balance out any swing-states shift toward Romney in national polling.

After the conventions, the national polls will matter a little less and state-by-state polling give us a better idea of where the race is.  Of course, one has to be careful which of these polls one considers.  Gannett’s News-Press reported last night on a new poll that showed Romney up by 14 points in Florida — but just how reliable is the pollster?

Foster McCollum White & Associates, Baydoun Consulting and Douglas Fulmer & Associates, of Dearborn, Mich., questioned 1,503 likely Florida voters Friday and found Romney, the presumptive Republican presidential nominee, leading Obama 54%-40%. The poll has a margin of error of +/-2.53%.

Although recent Florida polls have been trending slightly in Romney’s direction (+2% and +1% in the most recent surveys), the jaw-dropping 14-point gap is a shocker. Future polls will determine if this result is ahead of the curve or merely an outlier.

In the U.S. Senate race in Florida, the poll found Rep. Connie Mack IV, R-Fort Myers, leading incumbent Democrat Bill Nelson 51%-43%.

We got a few e-mails on this, but I’m very skeptical about these results.  The current RCP average for Florida is Obama +1, even though the last two polls listed show Romney with an edge.  The latest RCP average for the Florida Senate race has Nelson up by 2.5 points, not down by eight.  The last pollster to show Mack with a lead at all was Survey USA in July, by six points.  Otherwise, the average would have Nelson up by 5.3 points.  Finally, I have no idea who paid for three consultant shops in Michigan to combine up for a poll in Florida, but I wouldn’t be surprised if the answer had something to do with the outlier it produced.  I sent an e-mail to the News-Press digital editor to see if they have more data on this poll, and I’ll update this post with any reply I receive.

I’d stick with Nate’s analysis on this, and be mildly encouraged by the direction of swing-state polling this summer.  If nothing else, it shows that Obama remains a very weak incumbent, and until he starts talking about his second-term plans, swing-state voters will probably continue to drift to the candidate who has an agenda.