Watch the Polls: They have an important message

By John Zogby

"How can Americans say Bill Clinton is doing a good job as President when no one bothered to call me?" That’s a typical call we receive at Zogby International these days.

"I talk to 20 or 30 people a day and everybody hates Clinton", our frequent callers and e-mailers tell us.

"Just who are you calling anyway?"

I have also been fielding a number of calls from newspapers, talk radio hosts, and television news directors about polls in the midst of the controversy swirling around the President. I thought I would take this opportunity to try to clear up some confusion.

Contrary to any rumors that have been widely circulated, we pollsters do not focus most or many of our national calls in New York and California. Actually, we use sampling that is "random" and "stratified" which means that every household in the US must have the same chance of being called as every other and that we take special pains to ensure that every state and region are represented equitably.

None of us who are public pollsters -- i.e. polling for major media – wittingly produce polls that are skewed toward Democrats or Republicans. While I do have some disagreements with some of my colleagues about the oversampling of Democrats (simply because they are more likely to respond to polls than Republicans) this is a sampling issue and not the result of any built in bias or prejudice.

My firm only polls "likely voters" on matters of politics and public policy because they are the ones who actually count on these matters. Because of that, our polls tend to show less support for the President and Democrats in general because actual voters tend to include fewer minorities and lower income groups than all adults.

By and large, all of us do a good job of representing the mood and behavior of Americans. As with good consumers dealing with retail items and HMOs, good citizens must also be good consumers and evaluate for themselves which polls they trust.

All of this leads to a very basic question: why conduct polls in the first place?

In a democracy, public opinion must be a factor in any policy discussion. I personally have some trouble when polls drive policy or decisions by our leaders, but the opinion of voters must be somewhere in the mix. Ultimately, we elect our representatives to make decisions on principle and conscience, but we also expect that they not be contemptuous of the people who elected them.

How better to find out what the people are thinking than to do it scientifically? Make no mistake about it: polls that are done well can be an extremely accurate form of tracking public opinion. But there are two myths about polling that need to be cleared up:

Myth 1 -- Polls predict the winners and losers. Actually, a poll is only a snapshot of a moment in time. It can point to trends, but things can change on election day, when a lot of undecided voters make up their minds. We do try to ask "projective questions" – i.e. to see how people will react to situations and messages, but a poll can only measure a fixed moment in time.

Myth 2 -- Polls determine the outcome of an election (therefore, why bother to vote?). Polls generally only confirm what professional observers (and many voters themselves) already know – whether a race is close or not. Early polls can have an effect on a candidate’s ability to raise funds, but they do not shape how an election will turn.