By Shuting Wang (North Carolina State University)
Information
This hands-on, in-class activity allows students to apply theoretical survey sampling concepts to a real-world application. Students gain experience acting like statisticians working with government data. The instructor gatekeeps a population dataset containing the total number of workers and unemployed workers for all 100 counties in North Carolina. Students work in groups to design a sampling plan under the constraint that they can obtain data from only 10 counties. They then request data for their selected counties from the instructor (with support from a TA) and use this sample to estimate the unemployment rate in North Carolina, including a standard error for their estimate. Students experience joy and discovery through this engaging activity by making decisions and carrying out calculations themselves using real, socially relevant data. This context helps spark motivation, while deeper engagement emerges as students realize there is no single “correct” sampling plan and compare their approaches and results with other groups. There is also an inherent sense of joy and suspense when students request data for their chosen counties, mimicking a real “fieldwork” scenario. This activity is implemented at a large public university in an undergraduate survey sampling course for statistics majors and minors, with a class size of about 50 students. The structure of the course, along with TA support, allows students to collaboratively design sampling plans, engage with authentic data, and experience the decision-making and uncertainty inherent in real statistical work.