Opinion polls are generally based on knowledge developed through asking a number of people a question or a group of questions. The way that such surveys are conducted frames the results. The chosen methodology can have a major impact on the outcome, especially if a person or group hopes to influence the results.
It is important to know things like what questions were asked. In what order were the questions or possible answers given? Were the questions slanted toward a specific outcome, or impartial? The sample size is also important because the larger the samples are, the more accurate opinion polls tend to be.
When polling a particular group of people, it is necessary that this is made clear. If only women were polled, one cannot accurately claim that the results are representative of all adults. If only people selected from certain areas, age groups, or other demographics are surveyed, the results should not be accepted as accurately portraying how all Americans feel about a certain issue.
It helps to know who conducts opinion polls as well. Did a company pay to have a poll done regarding its product? Independent polling by experienced entities tends to achieve results that are more precise. A new or unknown entity does not have the same level of experience in designing opinion polls or in interpreting the results, but at the same time, a recognized name does not guarantee the results can be trusted. In some cases, the methods used may be outdated or not sufficiently inclusive.
When it comes to elections, opinion polls can become even more complicated. It helps to know if all the candidates were included in the poll, if all names were offered in the same manner, and in what order they were listed. If a candidate was excluded, the poll should be discounted.
Even seemingly subtle things like listing the candidates in order that the perceived top tier candidates or frontrunners are listed first, can affect the outcome. Listing candidates randomly tends to provide more accurate results. Listing names in a different way than other candidates are listed certainly affects the outcome. So, if you have to select “other” instead of your preferred candidate’s name, take the results of such opinion polls with a grain of salt.
Opinion polls paid for by a candidate tend to be suspect. Remember too that when only registered Democrats and Republicans are surveyed, such opinion polls will not be as accurate as those that include independents and third party voters. New voters may not be included either since lists are often compiled from previous election data. If a poll includes only previous voters then coming of age voters aren’t taken into account.
Events can also have an impact on the results of opinion polls. If a wedge issue suddenly gets a lot of play, if a candidate commits a faux pas, or even if something positive occurs, polls taken directly after an event can produce very different results. It is also true that opinion polls tend to become more accurate closer to an election. Early polls are not much more than popularity contests where respondents tend to choose household names. After candidates appear in debates and begin receiving press coverage, results can change a great deal.