More evidence cell phone users affect polls

The latest political polling numbers suggest that poll results sway in favor of one candidate or another depending on whether a pollster calls likely voters on cell phones, Nate Silver pointed out on FiveThirtyEight.com on Sunday.

(Credit: FiveThirtyEight.com)

Silver's analysis shows that out of 14 polling organizations, the five that call likely voters' cell phones (shown in gold in the accompanying chart) put Democrat Barack Obama ahead of Republican John McCain by an average of 9.4 points. By contrast, the eight pollsters that do not call cell phones have Obama ahead by 5.1 points.

The data mirrors studies that suggest likely voters who only use cell phones are more likely to vote for Obama. The Pew Research Center last month released a study showing that cell phone users favor Obama more than landline users by at least 10 points.

Cell phone users have presented a challenge to pollsters because they may no longer live in the state to which their phone number corresponds, and the cost of using cell phone minutes could affect their participation in polls.

[MProsch]

Are you joking? I don't know anyone that only uses their cell phone...as their only phone...that would vote for Obama!!!

[box-builder]

You know one now! This cell phone only user is voting for Obama.

[vernmax]

Make that two. Were you joking??

[discern]

M Prosch must be joking. Add another cell phone only Obama supporter.

[ry_jones]

well, I voted for Palin. Thanks.

[rcrusoe]

For the past 8 years or so the only polls that were accurate were the ones taken the day before the election. It will be interesting to look at this data on Wednesday.

[Pete Bardo]

MProsch, of course you don't know anyone "only uses their cell phone" that would vote for Obama, but that's probably a fairly small sample from which to draw your conclusions. We all tend to "know" people who walk, talk, dress and think the way we do. That's what makes them our friends! And this isn't about people who only use their cell phones--it's about whether the pollster called a cell phone or a land line.

What seems to allude most of us is the fact that we are "not normal". That's just a fact of life. Personally, any (and every) candidate who called my cell phone will not get my vote. They can spend as much of their own money as they like, but stop spending my minutes for me! It's bad enough they get to decide how to spend my tax dollars.

If one of the pollsters were to call my cell, I'd probably hang up. I wonder how that skews the results?

[E B]

The problem is, nobody REALLY knows how it scews the results. The polls all make guesses and adjust their numbers according to those guesses, but the polls depend on the accuracy of their guesses almost as much as they depend on the people who answer the phones.

Are newly-registered voters more likely to register this year than in the past? At what rate? What about young voters? Old voters? Are blacks going to turn out in higher-than-historic numbers because of Obama? Will hispanics?

Are cell phone users who hang up on pollsters going to answer more like other cell phone users, or more like the general population (landline users)? Are you getting a geographically representative sample when you call cell phones? How can you tell?

Are supporters of one candidate or the other more likely to respond to a poll? If people refuse to answer, do you ignore those numbers or call them undecided? Will people admit they aren't going to vote for Obama, when the clear implication in the media for the last year has been that only racists would vote against him? Do the questions in any way imply that you would have to be racist to vote against him? Might the questions favor one side over the other or inspire a partisan reaction, in any way, even in ways that seem unreasonable to a neutral observer?

There is a science to polling, and there is an art, and they intersect. There's a lot of interpretation that goes on, and those who are best at it have the strongest polls -- but this year, there's a lot of head-scratching as they try to figure out if historical norms are good models or bad ones.

[discern]

I won't be swayed by a candidate calling my cell. I probably won't answer if I don't recognize the number during this season. By the way, I think you meant to use the word "elude" in that context, and not "allude."

[ry_jones]

As a cell-phone only person for many years, I've (obviously) not been polled. The bit that interests me here, though, is that anyone still pays for minutes. I pay for something like 2100 minutes a month for my five lines; losing 10 or 20 minutes to a poll is epsilon. Does any post-paid user really care that much about minutes?

[bfades]

I've worked as a caller at a political polling place for the last year and a half. Most people get angry at you when you call them on their cell phone. Some take it as a personal attack on their privacy. Many yell about having to pay for the minutes and very, very few of them participate in the poll (and the ones who do only do it because the phone is paid for through their work.).

[DigitalFrog]

forget the cell phones, I hang up when they call my landline. I have no use for pollsters, there is only one poll that counts.