Tag Archives: bias in polls

Another Reason Obama May Be Doing Better Than You Think

Just wanted to point you to an article from last June in the Daily Kos exploring how the rise in cell phones may be effecting telephone polling. The theory is that because it’s more difficult to reach people who don’t have landlines, and because these people are often the young, or the poor, or minorities, pollsters may be missing a significant number of likely Obama voters.

This is a not a time for complacency, but there’s something to this.  I’ve been volunteer in Reading, PA at Organizing for America and have noticed that voter outreach by telephone is not what it was “back in the day” or even four years ago.  Most households, even those with landlines, don’t answer calls from numbers they aren’t familiar with.  While some of us “port” and protect our phone numbers (you’ll have to rip my 212 from my cold dead hands), many folks simply get new numbers every time they change phones.

Monday night I attempted to make 100 “voter turnout” calls based on a list of registered voters within a certain Reading neighborhood.  We were instructed to let the phone ring no more than four times, and not to leave messages on voicemail.  Most of the time, my calls were unanswered, but there were also quite a few disconnected numbers and even more “wrong numbers.”    That evening I did not reach a single voter on my list.  Yet, I know from canvassing that an overwhelming number of folks in the community support the President and have committed to voting for him.

If telephone pollsters are primarily reaching people who (1) have landlines, (2) answer their phones when they don’t know who is calling, and (3) are willing to answer a pollster’s questions, then they are not interviewing a representative sample, and come election day, this could provide a welcome surprise for Democrats.