While discussing the problems cell phones posed for pollsters in class today I stumbled across this Pew Research Center study on what they call the "cell only" problem. 13% of Americans only use cell phones and are more difficult to reach than those that use landlines. Since cell phone users are different demographically (younger, more likely to be Hispanic and African American, less likely to be married or a homeowner) this can bias results in away from those groups.
The study suggest that the difference at the moment is marginal, but they project that 25% of the population will be cell-only by the 2008 elections and that adjustments are necessary to ensure that the polls stay accurate. It's not impossible but it is more expensive. Reaching cell-onlies costs 4 to 5 times more than land line users.
