This week I devoted a class period to sampling techniques (random, systematic, cluster, stratified, and convenience). Students worked on an Interactive Reading Assignment before coming to class, and I began class with a quick discussion of the different types and their strengths/weaknesses. After the discussion I followed up with a class activity designed to help students understand how to select a sample using the systematic sampling technique as well as introduce them to some of the sampling features of StatCrunch. Most importantly – I wanted to demonstrate to my students that samples vary, a very important concept for them to understand moving forward in this class.

Click here to open a copy of the activity: Sampling Activity Fall 2016

I provided the students with a list of the genders of 76 students, numbered from 1 through 76. Together we worked through one sample of 10 students with a starting point of p = 1. I then had students repeat this process for different values of p and I wrote their outcomes on the board. We were able to see that the percentage of female students in the sample varied from 40% through 60%.

I then showed students how to create a list of random integers between 1 and 76 using StatCrunch. After we looked at my sample, students paired off to do the same. The results were more varied, but we did have approximately 30 samples instead of 7. This helped to drive home the idea that samples do vary. Although we only get to draw one sample, students need to know how samples vary in general.

For the last part of the activity, students joined my StatCrunch group and opened a file containing the responses of 76 students to a 3-question survey.

1. Are you a smoker?
2. Do you own an iPhone?
3. How much did you spend on books and supplies last semester?

Students had StatCrunch select a sample of 20 students and computer the percentage of students who were smokers, the percentage of students who own an iPhone, and the average (we haven’t learned to call it mean yet) costs for that sample of students. Again, we saw a wide range of results. I then showed them the results for the entire group of 76 students so we could think about how close our samples were to the actual measures of the 76 students.

If you would like to see the StatCrunch data set, use this link to the data set.

My students left the classroom with a solid understanding of sampling techniques and for the benefits of using StatCrunch.