Rasmussen’s latest (and last) finds Romney and Obama tied in Ohio:
The pivotal presidential state of Ohio remains all tied up on the eve of Election Day.
The final Election 2012 Rasmussen Reports survey of Likely Ohio Voters shows Mitt Romney and President Obama each earning 49% support. One percent (1%) favors some other candidate in the race, and another one percent (1%) is undecided. …
The race in Ohio was tied late last week after Romney posted a slight 50% to 48% advantage a few days earlier. The candidates have been within two percentage points of one another or less in every survey in Ohio since May.
Forty percent (40%) of likely voters in the Buckeye State have already voted. Obama leads 60% to 37% among these voters.
Today’s University of Cincinnati poll also found a statistical tie, with Obama up by one. Previously, Rasmussen showed Romney with a slight lead in Ohio. All of the other polls have found a tie or a slight edge for Obama — and several in the last week have found Obama at or above the 50-percent line. Party ID breakdowns that favor Obama could be skewing the polls, but we won’t know precisely to what extent until after the election.
If Michael Barone is right, the polling is wildly off-base and Romney is heading for a landslide:
Also, both national and target state polls show that independents, voters who don’t identify themselves as Democrats or Republicans, break for Romney.
That might not matter if Democrats outnumbered Republicans by 39 to 32 percent, as they did in the 2008 exit poll. But just about every indicator suggests that Republicans are more enthusiastic about voting — and about their candidate — than they were in 2008, and Democrats are less so.
That’s been apparent in early or absentee voting, in which Democrats trail their 2008 numbers in target states Virginia, Ohio, Iowa and Nevada. …
Bottom line: Romney 315, Obama 223. That sounds high for Romney. But he could drop Pennsylvania and Wisconsin and still win the election. Fundamentals.
On the other hand, if Phil Klein is right, Obama will prevail in all the states where he has a slight edge in the polls, and beat Romney by a hair:
I believe the arguments about polls understating Romney’s position have some merit, but only up to a point. I also believe that by and large, despite some high profile errors, polling is generally accurate when results from multiple pollsters overwhelmingly point in one direction. So, I’ve decided to split the difference in my prediction. That is, I’ve given Romney the states that are essentially tied, in which he’s led in at least some recent polls. But in states where Romney has trailed in nearly all polls, and in some cases by a comfortable margin, I’m giving them to Obama. My thinking is that even if Romney over-performs the polls somewhat, he still is unlikely to over-perform by a wide enough margin to win these states.
Applying this philosophy, I give Romney Florida, North Carolina, Virginia and Colorado. But I assume that Obama takes Nevada, Pennsylvania, Michigan, Wisconsin and Ohio. …
Add it all up and the final tally is Obama 277, Romney 261.
Klein has a safer bet, but Barone’s argument is persuasive. They both show how difficult it is to come up with an analysis when we know the polls have some flaws, but don’t yet know to what extent.