P1-04: Multi-institution Assessment Results from Courses Using and Not Using Simulation-based Inference

By Beth Chance, Soma Roy, Stephanie Mendoza, and Brannden Moss (Cal Poly, San Luis Obispo); Nathan Tintle (Dordt College); Todd Swanson and Jill VanderStoep (Hope College)


Cobb (2007) led a call for introductory statistics courses to center more on the scope and logic of statistical inference rather than mathematical theory. Several new textbooks (e.g., Tintle et al., Lock et al.) have restructured the content of the introductory course to introduce and reinforce inferential thinking using simulation-based approaches prior to asymptotic methods based on the Central Limit Theorem. In this poster we present some of the data from our multi-institution assessment project comparing introductory courses centered around simulation-based inference (SBI) to existing curricula examining both conceptual understanding and student attitudes towards statistics. We summarize results from a broad set of courses, spanning high schools, community colleges, four-year colleges and universities, including from many USCOTS attendees, in an attempt to identify possible impacts of different curricular approaches. Using hierarchical models, we explore the contribution of institutional, instructor, and student-level variables on student gains from the beginning to the end of the course. From these results, we provide recommendations for improving implementation of SBI (and non-SBI) courses and implications for second courses in statistics. Future data analysis will focus on retention six-months post course and a learning trajectory throughout a pure SBI course. These analyses will provide recommendations on how much of an SBI focus is necessary to improve student understanding and attitudes towards statistics. We are currently preparing cleaned sharable versions of the data.