Fairborn commencement
Badin graduation
Division I regional track finals
Ellen Belcher is the Dayton Daily News opinion pages editor. She writes about state government, education, the environment, higher education and all things Dayton.
Martin Gottlieb is an editorial writer and columnist for the Dayton Daily News opinion pages. He focuses on the political process itself and does such national issues as war, the economy, taxes and Social Security, as well as a hodge-podge of local and state issues.Home > Blogs > A Matter of Opinion > Archives > 2008 > September > 23
Tuesday, September 23, 2008
Are cell phones skewing election polling?
At age 40, I often describe my generation as the dividing line between those younger than me for whom technology — cellular phones, computers, e-mail, texting, the Internet, etc. — is a largely a “first language” and those my age and up for whom technology is more of a second language. That is, some of us are more fluent in having learned those technological communication tools than others. For the younger set, they basically learned technology naturally while growing up with it.
As someone who works at a newspaper, I like to point toward news consumption as a barometer. Those older than me tend to prefer to read their news on paper while those younger than me usually seem to follow the news online.
Cell phones are another way to see this dividing line. I haven’t quite been able to give up my land line at home entirely, but I’ve thought about it. A lot of my younger friends have long ago dropped the home phone and use a cell phone as their only means of voice communication. And it seems to me that the older you are the more likely you are to have a land-line phone.
If that’s true, it raises a question about election polling, which still is largely conducted using databases of exclusively land-line phones. The danger is that if the pool of those polled skews older because there are fewer young people with land-line phones, will the poll results reflected a point of view biased toward older voters? And are those points of view different enough from younger voters to give a false impression of the mood of the state or nation? We already know from other polling sources that older voters favor John McCain more and younger voters tend to lean toward Barack Obama.
One statistician has done the math and he thinks land-line only polls may be inflating McCain’s numbers by a couple of points nationally.
The analyst is Nate Silver at Five Thirty Eight an election projection blog. Silver makes no bones about his left-leaning politics, but he also is a respected statistician, known primarily for his work at Baseball Prospectus using mathematical formulas to project the performance of baseball players based on their prior stats.
Silver compared the results of a small number of polls that include cell phones to the majority of polls that do not and found a small lean toward McCain in the land-line only polls. Overall, he estimates traditional polls may be underestimating Obama’s support by 2 to 3 points. That is enough to make a big difference in the swing states that will decide the election.
After the election, there will be more data to analyze to nail down how predictive polls including cell phones were compared to traditional polls. And if there is a difference, it really could change polling as we know it. If land-line polls are really 2-3 points off the actual result, I expect those pollsters will change their protocols quickly.
Permalink | Comments (2) | Post your comment | Categories: National politics
