What do Coke and Pepsi have to do with statistics? The seniors in Matt Frankel’s interdisciplinary statistics class can tell you! After studying the statistical problem-solving process all year, students ran a large study with 300 subjects to see if they could differentiate between Coke and Pepsi with a greater accuracy than random guessing would provide.
 
The experiment was the culmination of students’ learning about the requirements of running a successful experiment. In this study, subjects received three cups filled with two kinds of cola and were asked to identify the one containing the different brand. The main goal of the study was to see if subjects could tell the difference between Coke and Pepsi with a greater accuracy than random guessing would provide with a one in three chance. Overall, the subjects were able to correctly identify the different cola with 46% accuracy. 

As any good researchers would, students had more questions after tallying the results of their findings and developed suggestions for improving the study for next year. Teacher Matt Frankel explains, “We were interested in knowing more about the subjects’ preference for Coke/Pepsi, whether they frequently consumed either beverage, or their confidence in guessing the different cup before/after. Since we only had time to run a single trial for each person, we thought that the influence of random chance may have been a part of our higher correct guess rate. We would consider running the experiment where we had three cups the same and one cup different (4 total cups) to reduce the chance of randomly guessing the correct cup to 25%.”

Overall, Mr. Frankel and his students considered their study a success and had a lot of fun in the process. “I hope students learn about the challenges of running an experiment and experience something logistically complex in action,” said Mr. Frankel about the study. “I want students to recognize that they are capable of being researchers and can be successful as mathematicians.”