Cell Phones and Polling

fivethirtyeight.com is one of the best poll geek sites on the net. Nate can parse the numbers in ways that put Chuck Todd to shame. Today, he attempts to answer one of the biggest questions facing pollsters and poll junkies in the entire election cycle: What effect are cell-phones having on the poll results?

Most pollsters don't call people on their cell-phones, but increasingly many people-- especially younger people-- don't have a landline. Does that mean that younger voters are being undersampled? Of course the pollster can always account for age by just polling more young people, but that might not be a good solution. For example, I suspect that young urban residents are less likely to have landlines than their suburban or rural counterparts, and at the same time, these are probably the most likely people to vote for Obama. The same is probably true of college students.

So today, Nate looked at the numbers, comparing the pollsters that call cell-phones vs. those that don't to see if there was a pattern. If the pollsters that called cellphones showed a consistent lean towards Obama, that would suggest that Obama supporters overall are being under reported by the polls. Nate's conclusion? Polls that don't include cellphones are shorting Obama by an average of 2.8%.

UPDATE: Be sure to read Nate's take on the Yahoo!, Knowledge Networks and Stanford University study on the effect that race will have on the election. The short summary? The study is flawed in ways that make its findings close to useless. Race may be a factor, but the study doesn't really help us know how big of a factor it will be in the end (or for that matter, whether it will be a factor against Obama or in his favor).