By Brianna Hitt (United States Air Force Academy) and Jessica Hauschild-Eastman (United States Air Force Academy)
Information
This session demonstrates two complementary approaches to alternative grading used in undergraduate statistics and data science courses that move beyond traditional points-based systems. In an applied statistical modeling course, students are assessed through a portfolio-based approach aligned with course learning objectives and descriptions of A–F work. Rather than accumulating points, students construct a portfolio demonstrating their learning throughout the semester. A key component involves selecting a statistical method not explicitly taught in class, independently learning it, and applying it to a novel dataset. Students then present their learning process, analysis, and interpretation in a conference with their instructor. To support growth and learning from mistakes, the course incorporates GenAI-supported exam reflections. After graded exams are returned, students engage in an optional dialogue with a GenAI tool, submitting their original responses, revising answers, and explaining misconceptions. The GenAI acts as a tutor or coach, helping students clarify misunderstandings and deepen conceptual understanding without generating solutions. In a core statistics-for-all course, a standards-based assessment framework centers on twelve core competencies students are expected to master by the semester’s end. Students demonstrate proficiency during graded reviews rather than accumulating points. Applied homework, projects, and specialty topics allow exploration while reinforcing statistical reasoning. GenAI-supported reading guides using NotebookLM let students interact with course materials, evaluate their understanding, and request quizzes or examples to engage with statistical ideas before and after class discussions. Both courses are offered at a military service academy with small class sizes (typically 15–25 students). The core course serves students from all majors, while the applied modeling course primarily serves mathematics, data science, and operations research majors. Both emphasize active learning, reflection, and applying statistical ideas to real-world data. This presentation will demonstrate how these grading systems function and how GenAI can support reflective learning. When integrated thoughtfully, AI tools act as partners in reflection, feedback, and exploration, augmenting rather than replacing student effort. By combining alternative grading with GenAI-supported activities, these courses foster curiosity, revision, and discovery. Students are encouraged to engage deeply with statistical ideas, revisit mistakes, and explore new methods, creating a learning environment where GenAI supports intellectual growth rather than shortcuts.