Don't Trust Any Poll That Omits Cellphones
They're ignoring 25% of the populace
By Kevin Spak,  Newser Staff
Posted May 24, 2010 1:33 PM CDT
Volunteers work in the phone bank Thursday, Oct. 30, 2008 at the McCain Victory Center in Lewis Center, Ohio.   (AP Photo/Kiichiro Sato)

(Newser) – Yeah, it's expensive, but pollsters need to start calling cellphones. According to the latest CDC data, 23% of US adults, or 25% of US households are cellphone-only, writes Nate Silver of FiveThirtyEight.com. Add in “cellphone-mostly” households and that jumps to 40%—and it could be even higher by November, never mind 2012. These people are just being ignored, even though they are very different from their landline counterparts.

Cellphone-only users are more likely to be young, urban, poor, and technologically savvy than their peers. “All of these characteristics are correlated with political viewpoints and voting behavior,” Silver argues. Pollsters try to compensate by “upweighting” the few young people they do reach, but that's a lousy solution, and a recent Pew study proves it. Pew pitted a weighted landline-only poll against one that called cells and discovered that the latter swung six points in Democrats' favor—a statistically significant result.