Student Notebook

Playing Guitar Hero to Understand Statistics

One day in my statistics class, a fellow student asked, “Can we play Guitar Hero in class? I think we could use it to teach something in stats.” This pitch did not immediately sell Michelle Verges, an Assistant Professor of psychology at Indiana University, South Bend, and professor of our course. So she responded, “We can play Guitar Hero if someone will write a brief proposal that explains how this game relates to your learning of statistics.” At this point, my mind was devoid of anything but designing this project.

At first blush, playing Guitar Hero, a video game in which players strum a guitar-shaped controller in time with real rock songs, seems like a student’s way of wasting time. However, after some contemplation, I realized that this game extends beyond entertainment by generating data quickly and presenting multiple conditions easily (sound on or off, distraction of the player, player order, etc.). With a little creativity, this game also gave our class the opportunity to improve our research skills, increase our confidence in understanding and applying statistics, and provide us with a realistic question to test. To illustrate, Guitar Hero is designed to calculate the average number of “notes” players accurately hit on a guitar controller. The guitar controller is similar to an actual guitar, except that it uses five colored buttons instead of guitar strings. In addition, the songs in this music video game are of finite length and number of notes, which can be used to create defined conditions. Guitar Hero allowed us to indicate the level of difficulty to reduce variability. Finally, this game produces nominal (e.g., sex), ordinal (e.g., performance rankings), and ratio-level data (e.g., accuracy). Taken together, these gaming functions were used to create an entertaining, hands-on demonstration to collect data for statistical analysis.

After writing up the proposal, I quickly set up an appointment with Verges to start designing the experiment. After a number of lengthy meetings,1 Verges and I decided that we would use Guitar Hero to demonstrate a 2 (sex: male, female) × 2 (condition: noisy, silent) × 2 (order: noisy condition first, silent condition first) mixed-factorial analysis of variance (ANOVA) in class. The noise condition was a within-subjects factor, whereas sex and order conditions were between-subjects factors. In short, we wanted to test whether the presence or absence of background noise would affect students’ performance playing Guitar Hero. We also wanted to examine whether there were any sex differences in performance.

I gathered, compiled, and formatted the data for analysis, and Verges created an SPSS assignment for students to individually analyze those data. The assignment evaluated our ability to identify variables, observe whether the assumptions of the ANOVA were met, report the results we obtained in the experiment, and provide an interpretation of the results. We were also asked to provide an open critique of the experiment based on the results. Typically, the SPSS assignments are quite difficult to complete, and this was no different. However, this time we were very excited to know the answer to “our” project, an idea that evolved from a sort of joke to a class assignment. Sadly, the project did not yield statistical significance because our analyses were underpowered. Though we failed to reject the null, we gained confidence and experience in our research skills.

The clearest outcome from my experiment was its positive impact on the overall outcome of the statistics course. I often hear fellow undergraduate students express their frustration with statistics, mostly complaining about the lack of application and practicality yielded, especially for students not intending to pursue a career in research. Thus, the opportunity to “liven things up” by playing a popular music video game in class was ideal. The Guitar Hero experiment demonstrated the utility of conducting student-directed class projects and the effectiveness of incorporating a nontraditional teaching approach to engage students in a complex subject like statistics.

The project also had positive outcomes for me personally. The mere fact that I developed a research project for the entire class to participate in has instilled a deep confidence in my understanding of statistics. I practiced theory, applied concepts, and engaged in research that is atypical for many undergraduates. Additionally, during the many hours working one-on-one with Verges, I was inspired to join APS, attend the APS 20th Annual Convention, become an APSSC campus representative (beginning my professional development), and discern my career goals within academia. This class project was an exceptionally rewarding process, and its benefits extend well beyond the classroom. Given my experiences, I encourage you to take such a chance if provided, be it a professor taking a chance to play a game in class or a student taking a chance to enjoy learning statistics.

1 As rewarding as it may be for a professor to work with a willing and able student, it is by far more rewarding as a student to work with a willing and able professor. The constant reminder of “What is the hypothesis?”, “What are my independent/dependant variables?”, and “How do we measure those variables?” provided by the professor was invaluable to me because of the instant feedback and guidance.

Observer Vol.22, No.9 November, 2009

Leave a comment below and continue the conversation.

Comments

Leave a comment.

Comments go live after a short delay. Thank you for contributing.

(required)

(required)