Campaign 2008

More
explainer
columns

What's a Bank Run?
And how do you get on the FDIC's secret problem list?
Jacob Leibenluft
posted July 18, 2008
So Help You, Dog
How does a canine cop become a "sworn officer?"
Brian Palmer
posted July 18, 2008
Blind Reading
Why would booksellers buy a title without knowing anything about it?
Noreen Malone
posted July 17, 2008
How Terrorists Say "Hello"
Do members of al-Qaida really give one another fist bumps?
Juliet Lapidos
posted July 15, 2008
Why Are Red Cross Reports Confidential?
The organization accuses the U.S. of torture but doesn't tell anyone.
Jacob Leibenluft
posted July 14, 2008
Search for more explainer articles
Subscribe to the explainer RSS feed
View our complete explainer archive

explainer: Answers to your questions about the news.

The Mobile-Phone PrimaryAre cell-phone users screwing up Ron Paul's poll numbers?

By Michelle Tsai
Posted Monday, June 25, 2007, at 7:06 PM ET

Download the MP3 audio version of this story here, or sign up for The Explainer's free daily podcast on iTunes.

Does everyone count?

According to a Gallup poll from early June, Rep. Ron Paul, R-Texas, is the preferred candidate of 1 percent of GOP voters. Paul's supporters, though, insist that the libertarian congressman's poor numbers are a result of pollsters undercounting voters who use cell phones rather than land lines. Are cell-phone users really screwing up the polls?

Probably not. Those in the cell-phone-only crowd tend to be younger, poorer, more liberal, and more likely to be an immigrant or minority. But there aren't enough of them today—about 13 percent of U.S. households—and they aren't different enough from other young voters to make more than a one- or two-point difference in overall poll numbers. In most polls, a one- or two-point difference falls within the margin of error.

The Pew Research Center tried to get to the bottom of this issue recently, conducting four surveys to compare cell-only and land-line respondents on 46 questions. On average, the answers of cell users and land-liners differed by 8 percentage points per question. The biggest difference between the groups: A paltry 49 percent of the cell-only cohort said they're registered to vote, while 78 percent of land-line users claimed to be registered. Thus, mobile voters might be underrepresented in surveys, but they also don't turn out on Election Day. Groups like MoveOn.org courted this set in 2004 but still found a smaller turnout than expected among young voters.

Cell-only supporters could give a boost to a candidate like Paul, who's at the bottom of the polls. An additional two points in the polls, though, wouldn't be nearly enough to move him to the next tier of candidates. Hypothetically, Barack Obama—who appeals to many college students and African-Americans—and Bill Richardson—who has lots of Hispanic supporters—might also find a few hidden points. But overall, top presidential contenders already draw such wide support across age, income, and ethnicity brackets that cell-only voters would have minor effects on their polling numbers. Then again, an extra 2 percent could mean victory in a tight election. Late-afternoon exit polls on Election Day in 2004 put John Kerry three points ahead of George W. Bush—just enough to give the wrong impression.

Pollsters usually can't reach the land-lineless today; they'd have to pay interviewees for lost cell-phone minutes, plus figure out how to get people to pick up when they're not driving or otherwise occupied. So, the polls try to compensate by giving more weight to answers from similar voters. That means they look for respondents who behave like the cell-only users. According to a CDC survey, such people are more likely to rent, be under 25 years of age, live with roommates who are not relatives, and live in poverty.

Cell-only constituents could actually become a less distinct political force as the group grows larger. (According to the Pew study, the size of the cell-only group could reach 25 percent by the end of 2008.) As more people ditch land lines, the cell-only subset could start to resemble the general population.

Got a question about today's news? Ask the Explainer.

Explainer thanks Mark Blumenthal of Pollster.com, Charles Franklin of the University of Wisconsin at Madison, and Scott Keeter of the Pew Research Center.