Marina Ptukhina (Whitman College), Julie Garai (Incredible Health)
Abstract
Students learning statistical modeling develop the skills to analyze data independently, but without an introduction to study design, they risk constructing misleading or incorrect models. This session introduces What Would Fisher Do (WWFD), a model-building tool designed to support students’ learning as they navigate increasingly complex scenarios. Using examples ranging from one-factor ANOVA to split plot experiments, we demonstrate how WWFD helps students transition from study descriptions to identifying sources of variation and, ultimately, to building models with clear implementation paths. Participants will leave the session with a set of worksheets designed to aid students in this process. This session is tailored for instructors teaching modeling courses, from introductory to advanced levels. Attendees are encouraged to bring writing utensils and an open mindset for collaborative work with fellow participants.