As an ASBA member, you now have free access to news and updates on important issues from our legislative team in Washington DC.

October 2006
by James C. Musser, ASBA Washington Representative

With the 2006 mid-term election just around the corner, public opinion polls of every stripe are filling the newspapers, the airways and the internet. The numbers often seem to be in conflict and it is hard to know which ones are accurate. There are, however, a number of key items to watch in evaluating polling data so that the reader has a better idea of what the poll actually means. The number of people polled matters and so does the manner in which the questions are asked and finally the methodology matters. Poll readers should also keep in mind that a poll is a mere snapshot of public opinion at a given moment and not a crystal ball that clearly shows the outcome of any particular race.

Size matters. Polls usually indicate how many people were surveyed and that is a crucial factor in knowing how reliable the poll is. Generally, the larger the number of individuals who are interviewed the more reliable the polling data is. For instance, if a poll surveyed over a thousand people, the likelihood is greater that the numbers in the poll accurately reflect public opinion than one that only interviewed two or three hundred individuals. The reader should always check the margin of error, which is clearly stated by reliable polls. The margin of error is generally smaller in large samples of one thousand or more individuals and higher in smaller surveys. The higher the margin of error the less reliable the poll.

How the questions are asked usually determines the answers. Good pollsters constantly struggle to refine the way in which they ask questions to remove bias. Good pollsters want to have accurate information so their clients really know what the public wants. Many good polls will even allow readers to see the exact questions that were used in taking the survey so readers can judge for themselves if there was any bias involved. A biased poll is not worth reading.

Finally, the methodology in how the poll was conducted is crucial in knowing if the information is accurate. Political polls that use a universe of “all adults” are less informative about the likely outcome of a political race than one that screens for “registered voters” and the most accurate are those that screen for “likely voters.” After all, on Election Day, the opinions that really matter come from those who actually voted. Another key point in looking at a poll’s methodology is how the respondents were chosen. The most accurate information is obtained by scientific random sampling that is careful not to include too many respondents of one party or the other, includes a proper balance of men and women and reflects an accurate or as nearly accurate as possible cross section of the voting age and ethnic populations. This part of the methodology is as much an art as it is a science and often accounts for the seemingly conflicting numbers in different polls on the same topic. Readers simply cannot rely on self-selecting polls such as those often found on the internet where one simply gives his or her opinion. These polls are subject to manipulation and do not necessarily reflect a broad cross section of the electorate.

These facts about polling should make the political races more interesting and allow poll readers to decide for themselves what the numbers really mean. Don’t forget to vote on November 7th. This is an important election for small business owners and all Americans.

James C. Musser, Esq. is a legislative consultant based in Falls Church, Virginia. His reports are updated monthly.