Tis the season to be wary — of polls

By JORDAN BRADLEY
Capital News Service
LANSING — Voters should cast a critical eye at polls before accepting their findings as truth, experts say.
Jeff Williams, chief executive officer of Public Sector Consultants, warns the public and media alike to be critical of polls this election season and the way they’re represented in the press. Public Sector Consultants is a private research and polling company based in Lansing.
Erika King, a professor of political science at Grand Valley State University, stressed the importance of knowing the source of a poll. She advised voters to look closely at what information is being provided, how the poll was taken, the kind of people being surveyed and how they were selected.

Here is what Williams suggests voters do:
• “Know your margin of error.”
It’s easy to get lost in the numbers.
If an article claims a lead by a small percentage, check the margin of error, Williams said. If the lead is within the margin of error, the race is technically tied, though the article may tell you otherwise.
For example, he cited a poll that showed Democrat Mark Schauer with 45 percent and Republican Gov. Rick Snyder at 42 percent with a margin of error of four points. The article reported that Schauer had a two-point lead. Though that may seem true, it was within the margin of error, technically showing a tie between the candidates.
Also, the public and journalists should know that accuracy is higher and the margin of error smaller the more people are polled.
Williams suggested focusing on polls with at least 600 responses, while King suggested looking at polls with at least 1,000 responses.
• “How confident are you?”
According to Williams, one in 20 polls is “hinky,” or inaccurate.
When viewing other polls, does one seem out of whack? If so, chances are that poll is an outlier.
King said voters should also look into the polling organization to make sure it is professional and accurate.
• “Timing is everything.”
Check the news for events that could sway people’s opinions. Was there a recent strike or scandal, for example? Was the poll conducted before or after the event? If so, opinions are likely to be drastically different.
• “Know your subject.”
Pay attention to the wording of questions. Does it sound as though it is leading to a certain answer? Keeping in mind that people’s perceptions are their realities, William said, do the questions play into perception rather than opinion based on the facts?
• “Know your polling universe.”
Is the poll conducted by a company or group that has a special interest in the results? If so, be extra wary. Polls from inside a party or business rather than polls conducted by news organizations are more likely to be biased.
King said, “One of the problems is that to save space in a newspaper, editors will include only a small description of the poll. They will almost never tell you how people are polled or sampled.” It’s crucial that pollsters include a diverse range of people.
“Most campaign polls reach people via telephone. What has to be done is to include cell phones as well as landlines. People who have landlines are different kinds of people, that is older, than people who only have cell phones,” she said.