Race, polling and why Obama might not win despite opinion surveys

Nothing commands an October headline like a presidential poll. Sunday's Review-Journal appeared to say it all: "Race to the Finish -- Nevada poll shows Obama pulling ahead as undecideds decide."

The newspaper-sponsored statewide telephone survey conducted Wednesday and Thursday by Mason-Dixon Polling & Research showed Obama passing McCain like he's standing still, 47 percent to 45 percent with 6 percent undecided. The racehorse news revealed Obama up from 39 percent in an August poll and McCain slipping from 46 percent.

Now numbers junkies can add a new statewide survey by Marvin Longabaugh of Magellan Research, which shows Obama ahead of McCain, 46.9 to 43.9 with 8 percent of the remainder either undecided or pleading ignorance.

Given the size of the national economic crisis and the unpopularity of President Bush, conventional wisdom dictates that McCain is gradually being sunk by forces beyond his control. Although each poll result falls within its margin of error, the surveys point toward an Obama victory in November.

Or do they?

Will Obama suffer from what some pollsters call the "Bradley effect?" That is, will he carry a lead in the polls until those remaining "undecided" voters are exposed as folks who would never vote for an African-American?

In Sunday's edition of The New York Times, pollsters and media researchers differed on the impact the Bradley effect might have on Election Day, but in a close race could the issue of race make a difference?

Mason-Dixon managing partner Brad Coker and Magellan's Longabaugh believe it's possible. The challenge is simple: Nevada political history is all but bereft of examples of African-American candidates running for statewide office in an era of blanket opinion polling.

"To say anything definite, you really can't because it's never been tested before," Coker says. "It's a factor, and you have to watch for it, but it's not going to show up anywhere but in the undecided column."

In more than two decades of polling, Coker says he's seen examples of the effect in several major campaigns, most of them in the South.

"What you have to caution against is what percentage of undecided voters are white," he observes. "If there are enough undecided white voters to cover the spread, you have to keep an eye on that."

In Coker's experience, "Undecided white voters typically broke strongly for the white candidate. It's never been tested at the national level. This will be sort of the first test of whether this effect has a national audience, or whether it's limited to places like the South."

But just because the Bradley effect exists doesn't mean it will make the difference.

While it appears to have been important in the case of former Los Angeles Mayor Tom Bradley, who lost the California governor's race despite late polls showing him with a sizable lead, Douglas Wilder of Virginia was able to prevail in a close gubernatorial campaign.

Longabaugh isn't so sure. He believes the race issues cut two ways. While some undecideds might be racist, he suspects others might be comfortable casting ballots for an African-American once they reach the privacy of the voting booth.

Would that be the "reverse Bradley effect?"

"We've never been through this before. I think it's always a possibility," Longabaugh says. "But certainly I think if we were going to see something like that we would have seen it in the primary. Maybe I'm naïve, but I don't think that's going to be an issue."

Measuring new voters and weighing their likelihood of turning out on Election Day, he believes, has greater meaning. Longabaugh broadened his definition of "inveterate voter" for his latest survey to include anyone who has cast a ballot in the past four years.

Likewise Coker is fine-tuning to capture voters whether young or old, inveterate or newly registered, and no matter if they use a land line or a cell phone.

But it's the racial issue that floats like an old ghost in '08. Will this be the year of its exorcism, or will larger factors such as the economy and the Iraq war prove the difference?

"If there is a hidden racial vote, it's going to benefit McCain, clearly," Coker says. "Is it there or not? Maybe in some states it is.

"Does it apply in Nevada? I have no idea."

John L. Smith's column appears Sunday, Tuesday, Wednesday and Friday. E-mail him at Smith@reviewjournal.com or call (702) 383-0295.