Sunday, February 10, 2013

Survey Says...

I really should get rid of my landline. The only time I ever answer it is when my mother calls. Most of the time, I don't recognize the number on caller I.D. and I just don't want to listen to a sales pitch or, worse yet, hear the voice of a pollster.

Because who are those guys? I don't trust them. When they say they aren't trying to sell me something, should I believe it? When they frame a question in a particular way, are they just trying to influence my answer?

Which makes me wonder if we can trust the results of public opinion surveys. George Gallup and Saul Rae believed democracy itself rests on the expression of public opinion. In Chapter 21 of our test, Gallup and Rae describe polls as a "swift and efficient method for legislators, educators, experts and editors to get a reliable measure of the pulse of democracy." I'm sure Gallup and Rae would be appalled at the Gallup poll released the day before the last election that predicted a win for Mitt Romney. Oops. Gallup has hired a polling expert from the University of Michigan to review the company's methods. (New York Times, Gallup Conducts Review of Tracking Poll Methodology)

Our readings last week also contained reservations about pubic opinion polls. Robert S. Lynd suggested counting noses doesn't necessarily add up to the truth and that companies like Gallup, "grow rich by perpetuating the public's sense of the competence of its opinions." (Chapter 21)

Certainly, contemporary pollsters are struggling with changes in technology that may influence the sampling of opinion. A study by the nonprofit Pew Research Center suggests the response rate to Pew surveys has dropped from 36% in 1997 to just 9% today. The study found that surveys that included both landlines and cell phones still did reasonably well in provided accurate information. But there were enough differences to raise eyebrows...and questions about survey reliability.

In a Slate article about the the reliability of surveys, author Will Oremus raises concern about various forms of bias and even suggests that public-opinion research was better suited to a time when polls were conducted by going door-to-door. That time is long past.

Oremus quotes census director Robert Groves as saying we're in the middle of a revolution when it comes to measuring public opinion and that tools such as internet data mining are being used to "fill in some of the gaps in traditional sample surveys."

"Will that revolution be the death of pollsters?" asks Oremus. "Survey says: too close to call."


  1. Looks like we were thinking along the same lines this week, Mary Jane! I hadn't read anyone else's posts before posting my blog tonight, but we are definitely in agreement. :)

    Since we both worked at 10/11, I bet we can also both agree that there is a certain type of person who also calls in to participate in a poll. There used to be a poll question on the 10 p.m. news when I worked there and there was always the same strange person who called in every night to give their vote. We always rolled our eyes and talked about how the sample was probably so biased!

  2. I know, I just read your post!
    No, I've never trusted those poll questions, either.