The “p-hacking-is-terrific” ocean - A cartoon for teaching statistics

Abstract

P-hacking is fishing for statistical significance through repeated testing on massive data. It would lead to spurious findings, misguide social practice and policy making, and thus should be avoided. Teaching about p-hacking is important, yet challenging. Cartoons are effective edutainment tools to engage students in learning statistical concepts. We created a cartoon and discussed how to use it in teaching about p-hacking by guiding students to think and answer a list of questions. This cartoon can be helpful with teaching both statistics courses and applied seminar courses in various other disciplines. Students are expected to gain a better understanding of multiple issues related to p-hacking, including its occurrence due to repeated testing, the problems with using an arbitrary threshold for the P-value and comparing statistical significance, the distinction between statistical vs scientific significance, the approach for interpreting testing results with a holistic view, and the strategies to avoid p-hacking.