Anything can happen in this election

Yesterday afternoon I suggested that Nate Silver of the New York Times was in denial- and perhaps engaging in a little wishful thinking- when he dismissed Mitt Romney's surprisingly substantial lead in the Gallup tracking poll in recent days (incidentally, today's poll still shows the Republican nominee up by six points over President Obama).

Matt Lewis of The Daily Caller agrees with me. But even so, maybe I spoke too soon. Ed Kilgore of Washington Monthly also cautions against taking the Gallup results too seriously, citing questionable methodology. And the fact is that most of the polls show a closer race. In fact, the Real Clear Politics national average shows the president leading by 0.1%.

It's hard to see how an election could be any closer than that.

Joe Battenfield of the Boston Herald sees Mr. Obama as being in big trouble. Several states in which the president should have no trouble at all- states like Michigan and Wisconsin and even Pennsylvania- have shown signs that they might actually be in play. And there is a great deal of confusion among the polls themselves, not only at the national but also at the state level.

The local news last night pointed out that only Wednesday an NBC/Wall Street Journal/ Marist poll here in Iowa showed POTUS leading by a full eight points. But a PPP poll out yesterday shows Romney leading by one. The first poll has a larger sample and a smaller margin of error,but the gap in results is simply too big for these things to account for the difference in outcomes.  Besides, the PPP poll is much closer to the mainstream of recent polling in this state; the RCP average has the president leading by 2.4% in Iowa.

While the polls do agree that Romney's favorable ratings have somehow overtaken the president's nationally, and seem to indicate that he's ahead right now in  North Carolina, and Virginia, and Florida, the consensus seems to be that Obama is slightly ahead in  Ohio.

The various estimates of where the Electoral College stands right now agree no more than the state polls do. Real Clear Politics, which bases its conclusion on the average of state polls, has Romney leading with 206 of the 270 electoral votes needed to win, and Obama with 201 and 66 up for grabs. Over at Race for 2012, they call it a 236-236 tie, also with 66 votes undecided.

But Election Projection- which allocates all the states to one candidate or the other- has the president winning, 281-257. Rasmussen calls it Obama: 237 - Romney: 235, with 66 electoral votes up for grabs. The New York Times' Nate Silver calls it 287.8 for Obama and 250.2 for Romney. How he manages to get fractional electors, I'm not sure. But Michael Barone of the Washington Examiner makes a convincing argument that in fact the president's electoral college pathway to 270 is narrowing dramatically.

I spent yesterday evening manning the phone banks at Romney headquarters in Urbandale. The Romney staffer who gave me a lift home had recently arrived from Boston. We had an interesting conversation about how the funereal mood at Romney national headquarters had been turned on its head by the governor's decisive victory in the first debate, and about the status of Iowa and the election generally. What he had to say pretty much corresponded with what the state party staffers seemed to think: that the election is now very winnable, both in Iowa and nationally- but that we should perhaps beware of getting too euphoric over what Gallup has been saying.

The fact of the matter is that the polls are all over the place. With a little over three weeks to go, either candidate could eke out a razor blade-thin victor- or blow the election open.

The only sure bet is that the next three weeks are going to be a whole lot of fun- and a whole lot of work- for those of us who are fascinated by presidential politics, and who happen to believe that a great deal depends on the nation's verdict on the night of November 6.

HT: Real Clear Politics
        Drudge