Facebook Friend Data: Analyzing Non-Random Samples in the Intro Course

• Activity Number: 610
• Type: Contributed
• Date/Time: Wednesday, August 7, 2013 : 2:00 PM to 3:50 PM
• Sponsor: Section on Statistical Education
• Author(s): Aimee Schwab*+
• Companies: University of Nebraska - Lincoln
• Keywords: Activity demo ; Facebook ; Nonrandom sampling ; Introductory course

Abstract

StatCrunch, an online stats app published by Pearson, includes a feature called StatCrunch Friend Data. Friend Data lets students sign in with their Facebook accounts, and pulls information from their friends' profiles. Undergraduate statistics courses don't spend much time talking about what happens to results with a non-random sample, and the Friend Data activity illustrates that! At the start of the activity, most students don't question the methodology (they're too excited to be on Facebook in class!). After they calculate their own confidence intervals, the students compare theirs to other students. Most immediately notice their intervals are extremely different, and usually very small. They also compare their intervals to published data about Facebook users as a whole. At the end of the activity, students discuss in groups ways to get a better sample from an online changing population - the general consensus is that it's tricky!

This activity keeps students engaged with a data set, gives them practice using a software program to calculate confidence intervals, and links back to earlier sampling and random variability concepts from the course.