Making Sense of Opinion Polls

Lee Atwater, the political consultant who helped guide Ronald Reagan to the White House, once famously told a reporter “perception is reality.” Take a look at the daily news cycle these days and you will realize how true his comment is.

The news is jam-packed with polls. Whether you are watching a nightly newscast, looking at the newspaper or browsing online news articles, it won’t take long to find out the percentage of Americans who believe a government policy is good or bad, the breakdown of people who follow a new diet plan, or even the percentage of working people between the ages of 45-65 who think they are saving enough for retirement.

And in 2014, the level of polling news dramatically increased because it’s an election year. In fact, depending on the findings of the poll, the survey itself can become the headline: “The Race is a Dead Heat!” or “Candidate X Pulls Ahead!” You get the idea.

The trouble is, there are so many polls conducted by such a wide variety of organizations, it is difficult to know just what to believe from all of the information that you are hearing. What’s more, polls are used to not only measure public opinion but also to shape public opinion. That’s why it pays to be a smart consumer of polling news.

So, with help from the folks at Elon University who survey North Carolina residents, here are a few thoughts to consider when reading about poll results:

  1. Who conducted the poll? Make sure the organization has a good track record. Media organizations and universities are one thing. A private polling firm contracted by a party or candidate should be viewed with a little more skepticism.
  2. What’s the methodology? In other words, how was the poll conducted? Did the pollster sample only telephone users with land lines? How many people were surveyed? Answers to all of those questions will influence the results.
  3. What’s the margin of error? Since pollsters only question a small number of people to reach a statistical sample of the population, the margin of error reflects the pollster’s confidence that the sample reflects the entire population. The more people who are polled, the smaller the margin of error will be.
  4. Look carefully at the question that was asked. Is it pretty straightforward or does it suggest an answer that caters to the organization sponsoring the poll?
  5. If the question doesn’t seem to make sense to you, chances are it was confusing to the people who answered the poll.
  6. When was the poll conducted? If it was close to a news event, such as a debate, an attack, or a disaster, chances are the responses to the poll will be distorted.


- Frank Graff

Frank Graff is a producer/reporter with UNC-TV, focusing on North Carolina Science Now, a weekly science series that airs Wednesdays, beginning in August 2013, as part of North Carolina Now on UNC-TV. In addition to producing these special segments, Frank will provide additional information related to his stories through this North Carolina Science Now Reporter's Blog!


Related Resources:


GSK