On Polls

Christopher Taylor has a well-thought out post on polls on his Word around the Net site. He discusses the history of poll-taking, how polls are conducted, and their margin of error. He also tackles the question "Why are polls sometimes wrong?" In answering this question, he focuses on the difficulties pollsters face in finding a representative sampling of the population and on psychological factors in play among the polled. Also of great importance for polling accuracy is the wording of questions in a neutral enough manner so as not to skew results. On the other hand, there are occasions in which certain groups may deliberately slant questions so as to shift public opinion rather than actually intending to report on opinions. Polls can also be a ruse to enable opinion to be reported as news. Additionally, Mr. Taylor discusses how campaigns tend become horse races with coverage focusing more on who is the front runner rather than on the issues.

Having had several communication, psychology, and statistics classes in college, I think I can say that Mr. Taylor is right on the money. I understand how difficult and how crucial it is to devise neutral and easy-to-use questionnaires. It is also necessary to obtain a fair sampling. When surveys are not well-designed or are misused, they become worthless.

If you get the chance, please check him out.