Sampling and Surveys (Lesson 4.3)
Chapter 4 - Day 4
Explain how undercoverage can lead to bias.
Explain how nonresponse can lead to bias.
Explain how other aspects of a sample survey can lead to bias.
Students have plenty of prior knowledge about how to survey well, even if they aren’t aware of it. For the most part, students can intuitively identify problems with surveys. Often, good sampling and survey methods are really just good common sense.
In the activity, students get three different examples of surveys that all have something wrong. Have students discuss in their groups what the issues are and what impact it will have on the results. We do not expect students to use the formal vocabulary in this discussion. We just want them to be thinking and reasoning about each context and the estimates being produced in those contexts.
This is where we can formalize student learning using the new vocabulary of undercoverage, nonresponse, and response bias. In the end, we want them to realize that all three of these issues can lead to estimates that might be lower or higher than the truth.
Some students confused nonresponse with voluntary response bias. Be sure to clarify with distinct examples of each.
Voluntary response bias is when an invitation to be part of a survey is given and people decided whether or not to participate. They were NOT specifically chosen by the researchers.
Nonresponse is when people ARE chosen by the researchers for the survey, but the people do not complete the survey.