Many would like to see public policy debate limited to facts, theories or even philosophies but perhaps nowhere else are emotions so prevalent in the discourse. To support their positions advocates often exaggerate and, if possible resort to studies and polls to prove their point. Poll results are often bandied about freely and just as often misinterpreted but the larger issue is the wording and sampling of the poll being quoted.

There is widely accepted methodology for ensuring that an accurate sample is obtained when taking a poll but professionals in this field readily admit there is room for error. For instance, young people and minorities are increasingly reliant on cell phones only for their voice communication. They don’t have a landline phone in their homes and most cell phone companies refuse to release the cell numbers of their customers. Most telephone polls have as their source for sampling the published phone numbers from directories published by phone companies. You can readily understand how this could affect the true validity of a poll sample. Be sure to see how large the sample is and what constraints were put on the people surveyed.

But an even more important factor in the validity of poll results is the wording of the questions. I was amused to read the latest results from the Civitas Poll, for example. The questions have been carefully worded and politically slanted to obtain results. When you read the questions it is no surprise they got the answers they wanted.

Here’s an example: “Currently North Carolina does not require voters to show a government issued photo ID to vote. The legislature is considering legislation to require all voters to show a photo ID when voting. Would requiring voters to show a photo ID improve the integrity and security of voting in North Carolina or have no effect on voting in North Carolina?” After such biased questioning I am surprised only 69 percent said it would improve security.

Just to hammer home the slant they followed up with question #10: “ The legislature is expected to pass a law requiring a government-issued photo identification, such as a driver’s license, passport or military ID, before being allowed to cast a vote in an election. They will also provide an ID free to anyone who can prove citizenship but cannot afford and ID. But Governor Perdue might veto this bill. With that in mind with the legislature who wants to pass this bill or Governor Perdue who wants to veto it?

But wait, there’s more. Another question follows up on the same topic. But then question #12 is a word for word repeat of question #10, but whereas the first time the results were 59 percent favoring the legislature this time the results for the same question are shown as 47 percent. And where Governor Perdue got support of 34 percent the first time the same question was asked she received 40 percent in the second. Which question are we to believe? More importantly, given the slant of the question should we believe either….or any of the results?

Our point is that you can make polls say what you want them to say if you phrase the questions so as to get the results you want. If you are going to believe poll results be sure to ask to see the questions and judge for yourself whether the questions are impartially written and easily understood. Like we say, when it comes to politics you are going to hear lies, damned lies and poll results.