What makes the pollsters' cell-phone-only problem interesting is not only how it renders election predictions this season suspect, but also how the problem reflects a real demographic shift in how people receive information and are mobilized to vote.

Not that I'm Marshall McLuhan or anything, but I often think that the ossification of certain boomer + media narratives (e.g. Obama = Carter, or 2004's re-fighting of the Vietnam War), have much to do with the opt-out of 18 to 35-year-olds from the television and print news media. 50% of those under 30 using cell phones only would seem to correlate well with people more likely to receive their news via the internet and less likely to receive their news on television. Just a hunch, but I bet it's true.

Anyway, another great piece on the cell-phone problem from the indispensable FiveThirtyEight.