In a democracy, politicians often put their fingers to the wind of public opinion when they decide what policies to advance. But people of all stripes are deeply skeptical of polling, especially when they think their opinions are moving in the wrong direction. Some of these suspicions are wrapped up in questions about how the pollsters do their jobs, like whether they manipulate the wording of questions or interview a biased sample. Other doubts are rooted in a deep distrust of the political parties, marketers and media giants that pay for these surveys.
The basic lesson is that the results of any poll are only a snapshot of the preferences, views and feelings of a group of people at a moment in time. Even the best polls are subject to all sorts of errors beyond their statistical margins for error.
These include sampling flaws, response instability, the tendency to over- or underreport certain kinds of behaviors and a lack of detail in how the poll is administered. These problems are different from – but can contribute to – the cold, hard statistical errors that are built into any opinion poll.
A good survey is also subject to the same kind of bias that you encounter in your own conversations and observations. Talking with your friends about an issue may give you a good idea of what they think, but these are self-selected samples that don’t necessarily represent the full range of experiences and viewpoints of a larger population.
About the Author
