McCain's Last Possible Path to Victory

The final round of state and national polling is in. It shows Barack Obama widening his average overall lead to 7.6 percent--a 2.5-percent increase from two weeks ago--and topping 52 percent nationwide. In the Electoral College, Obama is ahead by more than 5 points (again, on average) in enough states to reach 278 electoral votes; include the states where he leads by less--Florida, Virginia and Ohio--and that tally expands to 338. One hundred and eleven national polls have been taken since mid-September--and John McCain hasn't led in a single one of them. The prediction whizzes at FiveThirtyEight.com--who base their projections on polls--give the Arizona senator a 1.1 percent chance of victory.

But despite these statistical storm clouds, McCainiacs are still insisting that their man could win. Part of this, of course, is pure, unadulterated spin; no campaign admits it's going to lose before the polls close on Election Day. But there's also a glimmer of truth beneath the BS. The question for campaign reporters like me--and anxious voters of all political stripes--is whether there's enough reality here to actually upend the contest.

Basically, the McCain camp has been forced in the final days of this race to argue that everything we think we know--thanks, for the most part, to the incessant flood of polls--is wrong. In a memo released to reporters late last week, campaign pollster Bill McInturff--a reliable, no-nonsense guy--presented two poll-based pathways to a McCain victory. The first relied on undecideds. In this scenario, the vast majority of voters who persisted in telling pollsters until the last possible minute that they weren't sure which candidate they'd select on Election Day would, in the end, break overwhelmingly for McCain, propelling him past Obama and into the White House. "Given their demographics"--older, whiter, poorer Republicans, according to McInturff--"it is my sense these voters WILL vote in this election and WILL break decisively in our direction," he wrote. As a result, McCain's support would skyrocket--and Obama's would hold steady. Not particularly complicated.

McInturff's second pro-McCain scenario is a little craftier. That's because it relies not on voters who refused to commit but on those who "refused to even respond." As Pollster's Mark Blumenthal recently reported, "even the most rigorous national surveys struggle to achieve response rates over 30 percent"; the other 70 percent, meanwhile, hung up without participating. The question McInturff raised in his memo was, What if those who refused to be interviewed have very different political views--read: views far less favorable to Obama--than those who agreed to participate? He had reason to suspect that the answer was yes. In 1997, the Pew Research Center conducted an experiment that found "reluctant respondents significantly less sympathetic than amenable respondents toward African-Americans." Without these hard-to-reach anti-Obamans in the respondent poll, McInturff's thinking went, our national surveys would naturally skew toward the Democrat. But if they show up on Election Day, McCain would get a big--and potentially decisive--boost.

In other words, McInturff was claiming that McCain's two potential paths to victory were a.) the polls are incomplete or b.) they're completely and utterly inaccurate (mainly because they don't account for a bunch of anti-Obama voters who refuse to respond).

So what's the reality here?

Let's deal with "undecideds" first. Simply put, it's impossible to imagine that they will decide this race. The reasons are pretty clear. First, Obama has topped 50 percent in 28 out of the last national 29 polls, including all of the last 19; he averages 52.1 percent support overall. What's more, he crosses the magic 50 percent barrier in enough states (including Virginia) to earn him 291 electoral votes--21 more than the number necessary for victory. As a result, McCain could win over every single undecided voter--and he still wouldn't win the race. Never mind the fact that, far from breaking 100 percent for McCain, undecideds have split evenly between the two candidates in the final days of the race, with both the Democrat and the Republican's number ticking up 2 percent since late last week. Or that this "taking of sides" has left only 2 to 3 percent of the electorate undecided heading into Election Day--far too small a number to erase a lead as large as Obama's.

As for the "refused to respond" brigades ... well, they could conceivably still boost McCain. Earlier this week, Andrew Kohut, president of Pew, told Blumenthal that he's "always had a harder time completing interviews with a cohort of older, white, less well-educated respondents who typically demonstrate less tolerance on race-related questions." As a result, Pew identifies a demographic cohort that should be 29 percent of adults, but typically represents 20 to 25 percent of adults in their unweighted samples. While they always weight this group up to its appropriate level, he could not rule out the possibility that the missing respondents may be those with less racial tolerance, as they were in Pew's 1997 study." In other words, it's possible that 4 to 9 percent of the 2008 electorate will be exactly the sort of hard-to-reach anti-Obamans who refused to participate in the vast majority of pre-election poilling. Which would undoubtedly give McCain a bump.

Would the bump be big enough to install McCain in the Oval Office? Probably not. As Blumenthal reports, "Kohut remains uncertain about how much of problem these harder-to-reach respondents might be, he considers it something 'small to worry about' that might mean at most a percentage point or two in the results"--a boost that Obama's anticipated gains among new registrants and cell-phone users would probably cancel out. That said, what if that the "refused to respond" effect is larger in Ohio, Florida, Virginia and (especially) Pennsylvania than it is nationally--large enough, even, to swing those states in McCain's direction? At this point, it's a near impossibility. But it may be McCain's only hope.

Bottom line: if you wake up tomorrow to the phrase PRESIDENT MCCAIN plastered across the top of your local paper, blame the "refused to respond" brigades. And look for the headline NATION'S POLLSTERS COMMIT MASS SUICIDE somewhere nearby.

(NB: Even the McCain folks agree that the Bradley effect won't be a factor this year. In fact, race may help Obama more than it hurts him. I discussed why here.)