What are polls?
Polls and public opinion surveys are part of our everyday lives. From responding to surveys on social media to reading from the latest election poll, we are inundated with information about what the public thinks about a variety of topics and issues. Polling or surveying people means: “to ask people a question or a series of questions in order to get information about what most people think about something.” Polls are conducted by asking questions of a smaller set of people (or a sample) that represents the larger group. This is called sampling. The data is compiled and generalizations are made about that group of people’s opinions, based on the sample’s responses to the questions.
During elections, polls are used to share information with the general public and for candidates to gain insight into their standing in the race. During the 2016 Presidential election, as many as four polls a day were released in the months leading up to the election.
How are polls conducted?
Pollsters use “weighting” in order to adjust the results to match the population. For example, if a poll has a lower percentage of women respondents than there are in the U.S., the responses from that group of people (in this case, women) would be adjusted, or weighted more heavily, to make up the difference. The number of people who are polled for elections varies, but in general, between 600-1000 people are polled. The number of people that pollsters have to contact in order to reach that amount is much higher because some people refuse to participate or may not meet the criteria for participating in the poll.
Can polls have bias?
Bias and stereotypes can impact how people answer poll questions. For example, the “Bradley Effect” is named after Los Angeles Mayor Tom Bradley, an African-American man who lost the 1982 governor’s race. Some white voters, while intending to vote for the white candidate, told pollsters they were undecided or likely to vote for the black candidate (Tom Bradley) because they didn’t want to appear biased. This theory can explain discrepancies between opinion polls and election outcomes when there is a white candidate and a candidate of color.
Another example is during the 2020 Presidential election Democratic primary which includes several women candidates. In the poll, 74% of respondents said they were comfortable with a woman president but only one-third of them believe their neighbors are comfortable with a female candidate. This could affect poll results.
Exit polling is used to understand how people voted and to make predictions about the election’s outcome. Exit polls ask people after they vote how they voted. They are also asked to share aspects of their identity (i.e. demographic information) in order to learn about the election's results.
10 and up
Questions to Start the Conversation
- Have you ever participated in a poll or survey? What was the poll(s) about? Did the results surprise you and if so, how?
- Have you ever created a poll or survey? What was that like?
- What poll results have you seen before? When you hear about the results of a poll, does it influence how you think about that topic? How so?
- What have you been reading about current presidential election polls? What do you think about the polls?
- Why do you think there are so many polls about elections? Do you think they are helpful?
Questions to Dig Deeper
(See the Additional Resources section for articles and information that address these questions.)
- Why do you think it’s so important we understand how different identity groups respond to polls? Why is representation an essential aspect of polling?
- How can the questions asked in a poll lead to biased results? What can be done to reduce the bias?
- How do you think polls influence elections?
Ask: What can we do to help? What individual and group actions can help make a difference?
- Think about something you want to know about your peers at school (or some other group of people) and create a poll to learn about their opinions as a group. Use Survey Monkey, Poll Everywhere or a paper survey to collect the results and then analyze it.
- Volunteer with a local or national election campaign to learn about and get involved in election polling as well as other aspects of the campaign.
- If you see a poll that you think has bias or you see bias in the way a poll is discussed in the media, contact that pollster company or media outlet and let them know what you think. Conversely, if you see a poll that is unbiased and fair, commend the company who created or reported on it.
- Representing the People: Diversity and Elections (ADL Lesson Plan)
- Census Question Controversy (ADL Lesson Plan)
- Voting Rights Then and Now (ADL Lesson Plan)
- Teaching about Elections