Every major event has its signature trait — Thanksgiving has turkey, baseball games have their first pitch — and political elections have polls.
It’s a staple of a political campaign season — following a big speech or major event, polling centers across the country get the voters’ reactions by asking them, among other things, who they support in the election. The results, broadcast by the media and analyzed by the campaigns, come in the form of opinion polls.
The following is a brief guide to the anatomy of public opinion polls.
What’s in a poll?
Paula O’Loughlin , a political science professor at the University’s Morris campus, said one of the most important things to look for in a poll is the sample — is it a random gathering of people? Is the size large enough to represent the public as a whole?
“A poll has to be random or it doesn’t work,” she said.
Other factors, such as margin of error, or simply looking at the questions asked by the pollsters, are important as well.
Another important polling feature to look at, O’Loughlin said, is how pollsters contacted the participants. For example, telephone polls conducted over landlines often overlook younger voters who use cell phones, she said.
Rob Daves , a University teaching specialist who ran the Star Tribune’s Minnesota Poll for 20 years and taught a course on polling this past spring, said one of the most reliable polling methods still lies in telephone polls. New survey methods, like Internet polls, are still unproven in the industry.
He said studying the intricacies of how the poll was conducted is important for even casual observers.
“Most people don’t look at it even that closely, but they really ought to as good political consumers,” he said.
Quality in quantity?
A new trend in poll analysis, O’Loughlin said, is to compile many different surveys looking at a certain political contest, like a race for a Senate seat.
Although she warned about making sure the surveys ask the same thing, O’Loughlin said it’s a good way to look at the big picture in a race.
“If they’re all basically saying the same thing, that’s what you should listen to,” she said.
Clay Richards , assistant director of the Polling institute at Quinnipiac University , said a more reliable way to analyze poll numbers comes when compiling similar polls over time.
Poll numbers, Richards said, represent only current feelings. To get a good grip on how attitudes change over the course of an election, its best to look at how the poll numbers progress.
“In a fast moving presidential campaign, that can change,” he said, “change overnight, change within a day, change within the course of [Republican Vice Presidential Nominee] Sarah Palin’ s speech.”
Making a name
Daves said in the 2004 presidential election, his Minnesota Poll predicted that John Kerry held a four-point lead in the state.
During the election, Kerry won the state by three points .
“It doesn’t get much closer than that,” he said.
Daves said polling groups need to have a history of accuracy and validity before they can be truly reputable names in the polling industry.
“I think there is a lot more reliance on polling in campaign reporting now then there was a long time ago,” he said.
Quinnipiac has been in the polling business for 10 years and has expanded from covering Connecticut races to covering presidential races nationwide, Richards said.
A series of correct predictions in the 2008 primaries gave Quinnipiac national recognition this election year, he said.
“When the media looked at that, they said, ‘here’s someone we can count on,’ ” he said.
O’Loughlin said good polling can be a strong indicator of how people feel about politics.
“They’re quick and easy barometers of what the public thinks,” she said.