Share this article

print logo

Key indicators put Obama in driver's seat With 2 weeks to go, McCain finds 'bad news' across map

Barack Obama and John McCain are on a tour of unlikely battleground states that's proof of what the pollsters, the prognosticators and the pundits all say: The Democratic candidate is now the prohibitive favorite to be elected president.

Two weeks and two days before voters go to the polls, Obama found himself in North Carolina on Sunday, while McCain visits Missouri today. Both are states that President Bush won handily.

It all means Obama has expanded the political map in hopes of a blowout victory, while McCain is playing defense, trying to hold onto the states that Bush won in order to eke out a victory.

And it's a tall order, despite some tightening in the national polls over the weekend. Most importantly, polls in the key swing states that Bush won -- including Ohio and Florida -- have shown Obama with narrow but durable leads in the last two weeks.

"The bottom line is, in the last two presidential elections, the voter percentages barely moved over the last three weeks," said Charles Franklin, a University of Wisconsin political scientist and co-founder of the Web site Pollster.com. "That's very bad news for John McCain."

Right now, the numbers look daunting for McCain, the Republican nominee, especially in the state-by-state fight for the 270 electoral votes required for victory.

The polls show him losing by percentages outside the margin of error in several key states that Bush won: Colorado, Virginia, Iowa and New Mexico. Meanwhile, he trails by narrow margins in several other Bush states, most notably Ohio, Florida, Nevada, North Carolina and Missouri.

As a result, five popular election prognosticators show that, if the election had been held over the weekend, Obama would have captured as many as 364 electoral votes, in what would be an Electoral College landslide.

"If McCain is going to have any chance at all, it's a strategy that goes for 270 electoral votes and nothing more," Franklin said.

That means a laserlike focus on the swing states that Bush won where McCain remains in a tight race, most notably Ohio and Florida, Franklin said.

If McCain wins there, stages comebacks in Virginia, North Carolina, Missouri and Colorado, and wins one delegate in Maine's most conservative congressional district, he will be at the magic number of 270 without a vote to spare.

The trouble is, that's a lot of states to target for a candidate with limited time for personal appearances and a limited campaign budget that pales in comparison to his opponent's.

Of course, McCain could hope for a turn in the national trend, which shows Obama with a lead of about 5 percentage points.

But a big turnaround is not likely, said David W. Moore, a senior fellow at the Carsey Institute at the University of New Hampshire and the former Gallup pollster who wrote a book called "The Opinion Makers: An Insider Exposes the Truth Behind the Polls."

Presidential candidates with a lead as big as Obama's this late in the election almost always go on to win, Moore said.

>The Bradley effect?

Can McCain stage a comeback?

"There would have to be some sort of major October surprise that would alter the way the voters look at the election," Moore said.

But what if the polls are wrong?

It's possible, given that pollsters acknowledge a series of unusual circumstances this year that, in theory, could wreak havoc with poll results.

On closer examination, though, polling experts debunk the polling worries as nothing more than canards.

First and worst of all, there's the Bradley effect, or Wilder effect, named for African-American gubernatorial candidates in the 1980s whose final election margins were far different from what the polls predicted. That's presumably because voters told pollsters one thing about a black candidate while voting the opposite way in the privacy of the polling booth.

The Bradley effect would say that Obama, as an African-American, would lose a few points between the final round of polls and the election results -- but that's not what recent academic research predicts.

A recent study by Daniel J. Hopkins, a postdoctoral fellow at Harvard, found that the Bradley effect has not hurt black candidates since the mid-1990s. And in a study of this year's Democratic primaries, Hopkins found that polls accurately predicted Obama's share of the vote except in states with large black populations, where they underestimated his performance.

"I wouldn't be surprised if his actual rate of support is down slightly from the polls," but only because loyal Republicans who told pollsters they would vote for Obama will probably come home to McCain, Hopkins said.

>Youth vote called big

There's also much chatter about whether this year's polls are inaccurate because more and more people are ditching their land lines in favor of cell phones, which have numbers that many pollsters can't access.

But research shows that's not likely to make the polls inaccurate, either. The Pew Research Center for the People & the Press has conducted three polls since the primaries in which both cell phone numbers and land lines were called -- and there was only a slight difference indicating that cell phone users tend to favor Obama.

While pollsters disagree on exactly what is a "likely voter" -- and thus on who should be polled -- they're fairly confident that they are accurately predicting who will show up at the polls, despite the complication of huge gains in Democratic registration nationwide.

Obama, in particular, has a strong effort to turn out new voters.

"And that can really make a big difference in a close election," Moore said, noting that the election, at this point, doesn't appear to be close enough for it to make that kind of difference.

Even the poll with the most conservative definition of a likely voter -- one of Gallup's two surveys -- shows Obama with a narrow lead.

James E. Campbell, a political scientist at the University at Buffalo, plugged that Gallup data into his vote-prediction model last week and found a turnaround from last month's results. Based on the poll and the second quarter Gross National Product figures, Campbell's model shows Obama finishing with between 50.6 and 51.1 percent of the popular vote.

"I think one of the biggest things is the youth vote," said Campbell, who had previously been the only one of nine prominent political scientists with forecasting models who had predicted a McCain victory. "There's a bigger generational divide than I've seen in a long time," he said, with younger voters tilting heavily toward Obama.

In fact, McCain's hopes may depend on newly registered young voters not showing up at the polls, Campbell added.

Six of the other nine academic prognosticators show Obama garnering a majority vote of between 50.1 and 58.2 percent.

Two of those prognosticators call the race too close to call, but that's not how the gamblers see it.

Yes, you can bet on either Obama or McCain on Intrade, an Ireland-based prediction market. Sunday, Intrade put the odds of an Obama victory at 83.9 percent, just a fraction down from its all-time high of a few days earlier.

"There's a very close correlation between the fall of the Dow Jones industrial average and the rise of Obama," who was trading at 50 before last month's Wall Street crash, said Chad Rigetti, who serves as Intrade's vice president for business development.

"The chance of McCain winning is coming down to something extreme and extraordinary happening," Rigetti said. "Barring that, nothing is going to change."

e-mail: jzremski@buffnews.com

There are no comments - be the first to comment