The Future Of Polling

At 26 and 24, the more youthful members of the PIM team have not paid a landline phone bill since our sophomore years in college (2001-2002/2002-2003 respectively) and most likely will never will again. When we look through our contacts, we do not find one friend or peer whose number is a home landline. It seems the landline has gone the way of the typewriter, a relic of a bygone era.

Then why do so many political and social survey pollsters rely exclusively on landline phone numbers and their skewed demographics when conducting public opinion polls? They are cheaper and easier to poll, and the absence of this "cell-only" crowd does not really affect the outcome of most polls and surveys in a statistically relevant way, a 2007 Pew Research Center study found. The cell-only group at the moment is still small enough, and similar enough to its landline-owning demographic equivalents, to only affect surveys by .7% points on average, the Pew Research Center study found.

The 2006 study found that 11.8% of adults have a cell phone as their only phone line. Scott Keeter, one of the authors of the Pew study, estimates that the figure now is more like 14%. The cell-only population tends to be younger, more tech-savvy, less affluent, less likely to be married or a home owner, less conservative, and contain a greater proportion of men and minorities than the landline owning public. But it is growing fast, and will soon be a large enough population that traditional random digit dial (RDD) will not be able to accurately capture the true mood of the public. It would seem that excluding this group would greatly skew the results in races like the DFL Senate battle, where Al Franken and his youthful following would be under-represented against Mike Ciresi. Professor Paul Goren, a Ph.D. in Political Science at the University of Minnesota, assured us this is not the case. Data samples are often weighed following collection to compensate for a demographic that may be absent or disproportionately missing.

Pollsters, aware of the problem, have taken measures to court the ever-growing cell-only population, such as $10 to reimburse the time respondents spend burning their minutes. Outside of the inherent cost associated with cell phone dialing, further impediments arise. Federal law prohibits the use of automated dialing services when contacting cell users so each number must be dialed manually, and contact rates tend to be be much lower when dialing cell samples (how many people answer their phones when a strange number pops up?). The cost of the cell-only polls becomes more expensive, 2.4 times as costly as their landline brethren.

Keeter acknowledges the future of polling will look much different than it does now. In upcoming elections he expects Internet-based polling to become more popular, with access to the Web to be provided for those who lack a connection. There is also no word yet on the effects of new media like the social-networking site Facebook or the video-sharing YouTube on polls, but most experts agree the effect will be minimum at best.