'People-with-landlines' opinion polls?

In “A Powerful, Bold, and Unmeasurable Party,” George Gallup and Saul Rae state their case for the importance of public opinion polls. With The Gallup Organization having played such a major role in Lincoln and Omaha, it was interesting to read this selection from George Gallup. 

This article made me think about more about the validity of public opinion polls and the presidential election in 2012. As someone who does not have a landline in my home, I have never been asked to participate in a Gallup poll, or any other public opinion poll. As more and more people get rid of their landlines and only have cell phones, it makes me think that these polls should be renamed “people-with-landlines” opinion polls instead of “public” opinion polls. Is there a certain type of person who has a landline still? Probably so – namely, seniors/retired people or people with children of school-age – and therefore, not a representative national sample. I knew I had read articles last year about how President Obama was polling higher when both landlines and cell phones were polled, so I Googled this issue and found an article supporting this fact by Rebecca Rosen in The Atlantic from Sept. 20, 2012. Coincidentally, Rosen’s article made a connection between the landline versus cell phone issue in 2012 and the incorrect prediction made in the 1936 Roosevelt-Landon election that is described in the introduction to “A Powerful, Bold, and Unmeasurable Party.”

The introduction for the Gallup selection mentions the 1936 election in which Roosevelt won by a landslide but the well-known Literary Digest poll predicted that Landon would win, while Gallup predicted the election correctly. The Atlantic article from 2012 discusses how statisticians from the 1970s argued about how telephones biased the 1936 sample. A 1974 statistics book by Robert Reichard, The Figure Finaglers, said: “planners forgot one basic fact: the use of the phone itself was introducing a bias into the sample. Remember, this was 1935, and the people who owned phones at that time did not represent a cross section of the American public. Quite the contrary. Telephones were a luxury then -- and the people being sampled were the relatively affluent ones -- and hence the ones more likely to vote for the Republican candidate” (Rosen). In 1976, statistician Maurice C. Bryson argued in The American Statistician that “since voter participation tends to be highest among the well-to-do, the telephone owners shouldn't have been all that bad as a sample of the voting population.”

Rosen’s article brings to light a study done by Dominic Lusinchi in the spring of 2012 that was published in Social Science History about the 1936 election. Lusinchi applied contemporary statistical techniques to the Literary Digest data and found that, simply, non-response bias was to blame for the Digest’s prediction error. “Those 10 million ballots mailed? Only 2.3 million were returned. Landon supporters were far, far more likely to fill out the postcard and send it in” (Rosen). 

And, as Rosen says, while it’s unlikely that the landline versus cell phone issue will cause polls to be wildly incorrect these days, it is interesting to see how it does show the technological differences among generations of society.

On a related note, the incorrect presidential election prediction I always think of is the “Dewey defeats Truman” headline from the 1948 election. When I Googled Gallup and the 1948 election, I found that Gallup (along with pretty much every other public opinion poll) incorrectly predicted that Dewey would win. Gallup blamed the error on ending polling three weeks before the election. I was a bit surprised the introduction didn’t mention this fact.