Political polling is ubiquitous this election season, but such polls may be skewed because of how respondents are contacted. With 25 percent of American households only using using cell phones for telephone service (according to the CDC), surveys directed only at Americans using wired service may be subject to what experts call "landline bias".
Just how big is landline bias and how does it affect current election polling? The Pew Research Center attempted to find out. In a recent report, Pew estimated the bias this year is as large, or larger, than in 2008. The researchers compared 2010 Congressional polls aimed at both landline and cell phone users versus those targeting only landline users. The difference is striking. For polls conducted between Aug. 25-Sept. 6 among likely voters, the spread is: Republican preference 50 percent, Democrat preference 43 percent (landline and cell sample) and Republican 53, Democrat 41 percent (landline-only sample).
You can get a sense of the demographics of the cell-only crowd in this CDC report on the "wireless substitution," published in May. In general, younger adults are more likely to live in homes with no landline service. In addition, those more likely to go wireless are renters, the poor, and those living alone or in households comprised of unrelated people.
UPDATE: Field Research -- whose California Polls are often cited in The Bee -- samples both landline and cell populations in its surveys. That's because it takes random names from voter registration records where people are increasingly listing their cell phone numbers. Mark DiCamillo of Field Research says the percentage of cell numbers his group finds in voter records has risen from eight percent in 2005 to 17 percent in 2010.