Why NOT to Trust Polls

As we approach the elections of 2012 we will see a constant stream of polls. Many are reported by the media as if they were the absolute truth. Yet we should be asking if they are they being used to measure public opinion, or to influence it? Following are some of the reasons why we should be very weary of polls:

In many cases the organizations that conduct and report polls have a definite bias toward a desired outcome. This should be an immediate warning sign to everyone.

The U.S. Census reports that there are about 211 million voter eligible citizens, and 137 million registered voters. Typically polls are based on a sample size of about a thousand people. What this means is that they are sampling only 0.0000047 of the voter eligible population; or only 0.0000072 of registered voters, and they typically claim an error potential of plus or minus 3%. Oh yes, if you understand statistical sampling theory this would be perfectly adequate for sampling widgets off a production line, but it gets much, MUCH more complex when doing the non parametric sampling of human opinion.

Such things as how questions are asked...by letter, by phone, in person; how the question is worded and in what context; the series of questions asked along with the reported result on a given topic; the timing of the question vs. other happenings in the U.S. or the world; who is asking the questions can make a difference; if the question is asked by a person, voice inflections can make a difference...and so it goes. Sampling human opinion is tremendously complex. From personal experience, on polls reflecting critical issues affecting our country, I have asked the polling organization about some of the above questions raised above. I never received a response. Yet the polls spread like wildfire throughout the country and are treated as absolute truth.

David W. Moore a former senior editor of the Gallup Poll, now a senior fellow at the Carsey Institute at the University of New Hampshire published a book, entitled "The Opinion Makers". In his book Moore stated, "But the truth is that most of today's polls claiming to measure the public's preferences on policy matters or presidential candidates produce distorted and even false readings of public opinion that damage the democratic process." Does this statement from a former senior editor of Gallup give you pause on polls?

Mark Crispin Miller, a Professor of Media Studies at New York University commented on Moore's book, "A Powerful argument that polls do not merely misinform us but pose a genuine, if subtle, threat to our democracy." Robert W. McChesney, the Gutgsell Endowed Professor in the Department of Communication at the University of Illinois, commented "You will never regard political polls the same after reading David W. Moore's devastating inside account of their severe limitations and misapplications."

In many cases where public opinion seems to be headed in one direction, suddenly a poll will pop up, usually by an organization that has a definite bias, showing poll results giving confirmation of their bias in another direction.

For all the reasons given above, I do not trust polls. Should you?