Public Opinion Polls 2

November 6, 1998
Can I trust the polls? Are they statistically accurate? These are questions that many Americans are asking these days. Well, a recent article in the Washington Times says that it is possible to judge if a poll can be trusted. But you need to have sufficient information. And many print and broadcast polls eliminate dull but important details. Here's what to look for.

Who did the polling and who paid for it? If the poll was done by a political campaign, company, or trade group, the sponsor may have an axe to grind.

How many people were sampled? Generally the more persons interviewed, the more reliable the sample and the greater the accuracy. In a reliable public opinion poll, all in the target group have an equal chance of being selected. Public opinion polls where people select themselves to participate are unscientific and unreliable. That is also true of 900-number telephone polls and surveys of television show audiences or magazine readerships.

Who was interviewed? Surveys at times ask the same or quite similar questions of a sample of registered voters or of all citizens. Results will differ if they interviewed all citizens, registered voters, or those most likely to vote.

When was the survey done? Dates and times make a difference. Surveys done by phone are usually more contemporary than mail surveys which are slow and achieve low response rates. And man-in-the-street interviews might be amusing for Jay Leno's audience but they are typically unreliable.

Finally, what is the margin of error? There is always a degree of uncertainty, no matter how small. What is the margin of error? Typically it is about plus or minus three percent.

The next time to hear about a poll. Ask these questions. You'll be better informed.

I'm Kerby Anderson of Probe Ministries, and that's my opinion.

© 1998 Probe Ministries International