Opinion polls play a major role in business models, political strategies, public policies and the marketing industry. In its most basic form, opinion polling consists of pollsters asking members of the general public about their opinions on one or more specific subjects. Polling can take many forms, though many of the most thorough and ultimately useful polls fall under the classification of scientific polling.
Scientific polling is any polling that makes use of statistical information during the process of selecting participants. Before pollsters developed scientific polling, they typically surveyed members of the public at random or focused on specific types of poll participants to intentionally skew results.
Scientific polling uses demographic data, including gender, age, race, income level, geographic location, religion and political affiliation, to seek results that will be more accurate across a more broad population. For example, a poll that asks members of a multiethnic community which candidate they plan to support in an election can be scientific only if it has the correct percentage of participants from each ethnic group to match the percentages in the entire community.
The key advantage of scientific polls is that they account accurately for diverse demographics. Businesses, politicians and organizations may want to know how a specific sector of the community feels, or how the community in general will respond to poll questions. Scientific polling gives the option of focusing on a specific, targeted group, or expanding to include a representative sampling of the community. This means more accurate results and less bias on the part of pollsters because it reduces the possibility of human error due to prejudice.
Scientific polls are more complex to administer than random polls. Pollsters must first compile demographic data and then turn it into a model for administering a specific poll. The process of administering the poll is also more complex because it requires finding the appropriate participants and getting them to respond to a poll.
Compiling the results and breaking down the responses by each demographic group also takes more time, money and effort. Leaders who use poll data to make decisions have more numbers to consider when they review the results of a scientific poll.
Another weakness of scientific polls is their potential for inaccuracy despite extensive preparation and scientific analysis. Leaders who rely too much on data from scientific polling, or expect polls to be entirely accurate every time, can make unreasonable decisions based on limited or flawed scientific poll data. Scientific polls are costly to administer but can only be accurate when they reach enough participants. Details such as the wording of poll questions, the order of questions and the poll method (telephone, online, through the mail or in person) all can affect results. Pollsters include margin of error measurements to make up for some of the potential inaccuracies in a poll, but decisions based on scientific polls alone generally take on some risk.
When pollsters take steps to administer scientific polls fairly, and when analysts combine their results with common sense and other available data, they are useful tools in decision making processes. In particular, scientific polling can show how a group's attitudes and preferences change over time, as when the same poll yields different results on two separate occasions. Relative to random polling, scientific polling helps leaders make better decisions and engage with the members of their communities more readily.