Click here to download the PDF version.
Why Measure Public Opinion?
- Gauging public support on health issues is important to decision makers as they consider policy change.
- Learning the overall opinion of a population is one source of data for policy change.
Why Rely on Scientific Public Opinion Surveys?
- A scientific survey, or poll, is designed so that the results accurately reflect public opinion.
- Accurate results rely, in large part, on the method used to select people to complete the survey. The two most important elements of the method are random selection of the sample and weighting.
What does it Mean to Randomly Sample and Why is it Important?
- Random means that everyone within the designated group (i.e., city/county of interest) has an equal chance of being selected to participate. To keep the sample random and accurate, polling groups attempt to reach the selected participants multiple times.
- A random sample is the most accurate way to know a community’s true opinion of an issue without having to ask every person for an opinion. A non-random sample does not accurately predict the opinion of the city, county, or state as a whole.
- Polling companies typically select a sample of numbers generated by a computer using Random Digit Dialing. Sources, such as phone books, are not appropriate because they do not include all adults, may contain businesses, and do not account for people who have just moved in or out of the community.
What does it Mean to Weight a Sample and Why is it Important?
- A weighted sample for a public opinion poll means that the sample has undergone a statistical equation, called weighting that increases the accuracy of the sample to best represent the population. A weighted sample allows for the most equal representation of demographic characteristics such as age, gender, race, etc.
How do You Know if a Vendor Appropriately Conducts a Scientific Public Opinion Survey?
- The vendor should help you create a clear, reliable survey and then test the survey for you.
- They should select people randomly, and make 10-15 attempts to contact each person during various times of the day.
- They should be able to talk with you about how to randomly sample an appropriate number of people; this will depend on the size of the community you are trying to sample.
- They should be willing to weight the data if necessary.
- The vendor should be concerned with carefully and securely storing all data to protect the privacy of the participants.
After the Data have been Collected, What should be my Next Communication with the Vendor?
- Ask for statistics on the number of calls attempted and the response rate.
- Ask the vendor to report how the data were weighted.
- Carefully and securely store all data to protect the privacy of the participants.
For Additional Information about Scientific Public Opinion Polling:
- University of Kentucky, Survey Research Center: https://survey.research.uky.edu/#Anchor-Qualit-47087.
- Roper Center, Public Opinion Archives: https://www.ropercenter.uconn.edu/education/polling_fundamentals_intro.html