The Daily Tar Heel
Printing news. Raising hell. Since 1893.
Saturday, May 18, 2024 Newsletters Latest print issue

We keep you informed.

Help us keep going. Donate Today.
The Daily Tar Heel

Interpreting polls requires caution

As polling organizations across the country try to find the pulse of the nation, experts warn that the public should make sure they know how to read the numbers right.

Since the presidential race was solidified in March, pollsters have been turning out numbers on where likely voters stand. But simply relying on the figures isn't enough, and voters should be aware of what goes into polls and what they mean.

Still, "When used properly, (polls) certainly can be a snapshot of how a certain group of people feel at a certain time," said James Simon, a journalism professor at Fairfield University in Connecticut.

Frank Newport, editor-in-chief of the Gallup Poll, said the organization reports two figures on its presidential election surveys: the possible outcome if all registered voters were to cast ballots, and the outcome if all likely voters head to polling sites.

To find likely voters, Gallup asks people questions on a predetermined set of criteria. Respondents' answers then are judged by a computer model, and from those results, a pool of likely voters is put together based on the assumed turnout on Election Day.

"We're simply trying to say, 'Here's where it stands today,'" Newport said.

Likely voters include those who have voted in the last few elections and have been closely following the race, Simon said.

But new voters - like many college students - are treated separately. Oftentimes, such people get onto the list of likely voters thanks to their enthusiasm, which can be fleeting, said James Stimson, a UNC-Chapel Hill political science professor.

Pollsters also rely mostly on land-line telephone numbers to survey people. Voters who own only cell phones, including many young people, are not included, thus limiting the number of people available to question, said Robert Stevenson, a journalism professor at UNC.

Stevenson added that in general, people are increasingly resistant to answering questions. "I think as a general rule, the cooperation rate is dropping for all polls."

But deciphering which polls are reliable is difficult, particularly with so many online resources.

Simon said that the smaller the margin of error, the better. Most accurate polls have such margins no larger than 3 percent.

He added that the reputation of a poll also is important, and that people should be skeptical of any wild swings in numbers.

Stimson said sample size also is a good indicator, with 1,500 being a good size. But the public also should be careful in how the media report the polling information, he said.

Though the media work hard to be accurate, some take questions out of context or fail to compare the numbers they are using to others, which Stimson said is important in reporting numbers correctly.

But polling could soon become obsolete, as people screen calls and trade in land lines for cell phones, Stevenson said. "Those problems are looming larger and larger and probably at some point, maybe even in this election, the polls are going to get it wrong."

Contact the State & National Editor at stntdesk@unc.edu.

To get the day's news and headlines in your inbox each morning, sign up for our email newsletters.

Special Print Edition
The Daily Tar Heel's 2024 Graduation Guide