By Khanh Do (Grinnell College) and Joyce Gill (Grinnell College) and Shonda Kuiper (Grinnell College)
Information
Multiple regression is a cornerstone of statistical analysis, yet students often leave introductory and intermediate courses with persistent misconceptions about how to interpret results. This beyond session presents a freely available online lab activity using real institutional data from the Integrated Postsecondary Education Data System (IPEDS) to help students explore, diagnose, and correct common interpretation errors. By manipulating predictors, adding or removing variables, and visualizing coefficient changes in real time, learners see how model conclusions shift across different specifications. The app is designed for live classroom demonstrations or hands-on lab use and emphasizes conceptual understanding rather than memorizing formulas. This lab activity was implemented with approximately 20 undergraduate students in an intermediate statistics course at a small liberal arts college in the Midwest. Students first explored the tool individually and then engaged in a guided discussion about model assumptions and interpretation. By transforming regression from a static formula into an interactive, exploratory experience, the app allows students to experiment with their own questions and observe how statistical conclusions change. In an era where AI can easily generate rudimentary models, it is critical for students to learn how to construct meaningful models and interpret them correctly. The activity demystifies predictive analytics and encourages students to critically evaluate models rather than blindly trusting outputs. Attendees will leave with practical examples of how to integrate this R Shiny lab activity into their courses, strategies to assess and support students’ conceptual understanding of multiple regression, and ideas for using interactive model manipulation to deepen understanding of bias, multicollinearity, residuals, and statistical interpretation. They will also see a template for creating reproducible, scalable interactive tools for classroom use.