I’m a little behind the times here. Please don’t hold it against me. This is a graph that shows Obama’s lead in various polls on November 2:
It’s courtesy of the great non-partisan poll aggregator fivethirtyeight.com. Here’s Nate Silver of fiverthirtyeight describing it:
The polls in the Cingular-y orange color include cellphones in their samples; the polls in gray do not. The cellphone polls have Obama ahead by an average of 9.4 points; the landline-only polls, 5.1 points.
This disparity is known among pollsters as the “cellphone effect,” “the cellphone problem,” or “cellphone bias,” and it represents one thing that pollsters are having trouble adjusting to lately about emerging demographics – it’s taking a while for pollsters to realize the differences between landline-only and cellphone-only voters, and incorporate those differences into either their respondent distribution or their weighting strategies. While there are correlations between cellphone use and factors like race and sex, the most obvious correlation is with age; this correlation clearly comes through in the bias of cellphone polls toward Obama. Here’s a similar table from Pew:
While different pollsters come up with different numbers describing how much of an error this effect usually introduces into their outcomes, the error usually ranges between 1 and 3 points. According to Pew,
These problems are all the more pressing as the number of Americans who are reachable only by cell phones increases. U.S. government surveys estimated that about 15% of adults were “cell only” in the fall of 2007 and the rate of increase since 2004 has been at least 2% a year, meaning that the number may be as high as 17% in this election cycle.