iOS app Android app More

John Wihbey

John Wihbey

Posted: October 1, 2008 05:35 PM

The Cell Phone Polling Gap: An Artificially "Close" Race?


Despite Barack Obama's recent surge in the polls, the media maintain -- as they have throughout the summer -- that the race is close, and will be so on Nov. 4. And that may be true.

But imagine if Obama's upward climb in opinion polls right now were not just a few points. Imagine if it were seven or eight or more.

The media narrative might be clear: unfolding landslide. With that would come the momentum, the fundraising potential, the kind of late-election heat on the John McCain-Sarah Palin ticket that can induce panic.

And given recent insights into polling data and cell phone users, it's entirely possible that the only thing between a decisive Obama lead coming into October and more election-as-nail-biter boilerplate is the vast leftwing wireless network.

The Pew Research Center's recent report on the issue asserted that polling by landline telephone may undercount Barack Obama voters by perhaps 3 percent. It's one of the first pieces of conclusive evidence on the issue. But this is a young science, in a rapidly changing communications landscape where mobile-only households are multiplying. So who knows if that's undercounting the undercounting?

Statistically, Obama voters are more likely to be part of the younger, cell-only set. Cell users under 30 go left in a big way: 62 percent Democrat to 28 Republican. And cell-only voters of all ages go for Obama over McCain by 19 percent, 55 to 36 percent, according to Pew's most recent survey.

However, the Pew report points out that previous weighting techniques by pollsters assume that cell and landline users are the same politically. So many landline polls this year may have relied on faulty math.

A few poll organizations, such as Gallup and Pew itself, have been including cell phones in their surveys all year. Others, like NBC/Wall Street Journal and ABC/Washington Post, just started recently, according to Nate Silver at FiveThirtyEight.com.

Since the issue of a cell phone/landline gap first surfaced in the 2004 campaign, the problem of properly measuring opinion in a mobile age has been debated among pollsters and media wonks. The John Kerry campaign claimed its voters were being undercounted because some of them were cell-only, and exit polls confirmed that. But cell-only voters then represented a meager 7 percent.

Some states now have at least 16 percent of households reporting cell-only usage, more than twice the rates reported in the 2004 election. And that estimate is based on 2007 data. Many of the swing states in which the '08 race will be decided have high numbers of cell-only households: Virginia, Florida, Ohio, Michigan, among others. (See Brian Schaffner's discussion of cell phone-use and states at Pollster.com.) So, that could boost the case that the mobile phone gap could be playing a big role this year.

Of course, conventional wisdom holds that younger voters are less predictable as to whether they will actually show up to the polls, and the Pew report flags that issue. Young voters also won't represent the same force in a general election as they did in the Obama-mania primary, as their power is diluted by sheer numbers.

Yet the implications of the cell phone gap could stretch even deeper, perhaps into psychological territory. Everyone knows that '08 should favor Democrats and Obama, if history and issue-oriented surveys are any guide. And yet, as the mantra goes, the McCain-Obama race remains "close."

This is at a more speculative level, and this is where empirical data fails, but it's worth saying: a "close" campaign also feeds the notion that voters are just unsure or "undecided" about Obama -- whether on legitimate issues like his years in office, or other dubious ones.

And those undecided voters are now squarely in the spotlight. Their opinions on Obama's character may indeed decide this election.

Polling guru Mark Blumenthal recently dissected a group of 973 undecided voters and their views on whether McCain or Obama was "more prepared to lead the country." The result: McCain favored 52 to 18. For sure, "prepared" may mean just that -- ready to go, or seasoned by experience. But as has been endlessly discussed in this campaign, such vague language may also serve as a proxy for any number of other issues in voters' minds with regard to Obama. It's where a kind of subterranean logic can operate: a lot of other people are unsure about Obama, so I should be, too.

As David Moore demonstrates in his new book The Opinion Makers: An Insider Exposes the Truth Behind the Polls, surveys have a way of "manufacturing" a climate of opinion and shaping a public mood. The echo chamber of popular opinion has a way of reinforcing itself. (For a chilling anatomy of a polling nightmare, see Moore's first chapter on polls and the lead-up to the Iraq War.)

That is not to say that the cell phone polling gap is the only factor restraining Obama from coasting to a win. But with momentum at a premium in elections, and this one in the home stretch, it's worth considering that Obama's lead may be artificially dialed back by the demographics of his biggest fans.