The Use of Polls to Analyze Public Opinion in Politics
Public opinion is defined in the text as “the distribution of the population’s beliefs about politics and policy issues” (Edwards, Wattenberg, and Lineberry 150). On paper, it sounds so simple; in reality it is much more difficult to determine. The most common method for ascertaining and consolidating public opinion has been through the widespread use of polls. Their popularity has steadily increased over the years. One reason is that they provide an accurate, reliable representation of the opinions of an entire population and supply decision-makers with valuable insight that may be used to determine a future course of action. However, not all polls are created equal. Polls are conducted by various organizations, businesses, corporations and public officials in an effort to determine the public’s stand on issues ranging from the terribly controversial to the completely trivial. Regardless of subject matter, there are basic principles of polling that greatly affect their quality and reliability.
“Commandment #1” on everyone’s list states the necessity of a randomly selected sample of a population. This ensures that all those whose opinion the poll attempts to represent shall have an exactly equal chance of being interviewed. Telephone interviews conducted on a sample size of 1,000 – 2,000 people called from a list of random, computer-generated phone numbers are all typical components of a good poll. Some of these components can be altered without critically affecting the overall quality of the poll. For example, good polls can be conducted on a sampling as small as 700 – 1,000 people; the decrease in size causes an increase in error margin, but not by much. But the selection process should not be changed; random selection is essential to the goal of polling, which is to “come up with the same results that would nave been obtained had every single member of a population been interviewed” (Gallup 1). Without a randomly selected sample, the results can only be applied to the specific persons questioned (Blake 1). This is why results from polls that allow self-selection, such as those found on the web, or in your mailbox, automatically demand a higher level of scrutiny and skepticism.
Wording and ordering of interview questions is another area that demands major attention. Words, terms and phrases should be stated and ordered in such a way as to create as little room for bias as possible, and should be presented, in full, with the polling results (Blake 2). Question wording has been called the “biggest source of bias and error in data”, and is therefore an aspect of polling that will always leave room for criticism (Gallup 4). Occasional ill-wording of questions is tolerable, if not unavoidable, as long as it is not obviously meant to tilt the results or prompt a certain response.
However, it is not unheard of nor is it uncommon for such polling practices to occur. ...