Use of small-scale classroom experiments to inform simulation-based statistical inference (SBI) pedagogy in tertiary classrooms

By Sophia Chung, Maddie Schroth-Glanz, Beth Chance (Cal Poly - San Luis Obispo)


In order to help students understand statistical inference, we have been using different simulation strategies and representations of those strategies, e.g., PowerPoint animations, tactile/physical simulations, and computer-based simulations. We have also been investigating the effectiveness of each of these strategies on student understanding of key concepts. The general framework is the same for each strategy: carrying out repeated re-randomizations and building a “null” distribution. This research involves examining the impact of different levels of student involvement in creating this distribution (e.g., reading about creating the distribution vs. actively participating in creating the distribution). Students are randomly assigned to slightly different pre-lab assignments. For example, students may be asked to shuffle index cards and record results, or be asked to view a PowerPoint animation of the process, or be asked to perform a few trials in a free online applet. All students will then answer the same set of multiple-choice questions about the underlying statistical principles. Our prior suspicion is that direct student involvement may be preferable to “instructor demonstration,” but that is only based on opinion.