Mike Alger raises some important issues in his May 24 letter to the editor (“Smoking Poll Raises Questions about Accuracy of Results, Methods Used”). As he correctly notes, the article about the smoking poll conducted by our firm (“Ban Smoking in Restaurants, Majority Says”) does not give the reader enough information to evaluate the results of the poll. Let me explain how the poll was conducted and why we have confidence in the results.

Four hundred voters participated in the survey. How can we claim to know the opinions of all Fayette County voters based on only 400 interviews?

The answer is that we don't know exactly, but the process of “sampling” lets us make a pretty good estimate. Sampling sounds fancy but we do it all the time. For example, we decide whether or not to add salt to a pot of soup based on a small sample — a single spoonful. Because that spoonful of soup needs to be “representative” of the entire pot, we stir the pot before taking our sample. That gives us a random sample of the soup.

The fact that we surveyed a random sample of Fayette County voters is important, because it is the randomness of the sampling that allows us to claim that our poll is “representative.”

Even with what seems like a small sample of 400 voters, we are able to make a surprisingly accurate estimate of the attitudes of all voters. How accurate a poll is can be judged by its “margin of error.” In the case of our study, the margin of error is plus or minus 5 percent at a 95 percent confidence level.

What this means is that if we conducted our study 100 times, surveying 400 voters each time, 95 of those studies would give us results within five points of the true number. This, combined with the fact that we have had strikingly similar results when conducting the same poll in neighboring counties, allows us to be quite confident that the survey results we reported accurately reflect the attitudes and opinions of Fayette voters.

Beth S. Schapiro, Ph.D.

President

Beth Schapiro & Associates

Back to Opinion Home Page | Back to the top of the page