Graduate Student Research Symposium 2026

Generative AI (GenAI) is rapidly entering higher education, yet emerging “AI divides” may reflect not only access to tools but also unequal opportunities to learn how to use GenAI responsibly and effectively. Guided by an ecological perspective (macro/exo/meso/micro), this talk proposes a quantitative framework for examining how institutional policies, instructional practices, peer norms, and individual characteristics jointly predict students’ GenAI usage patterns. I will describe a feasible analytic plan using secondary survey microdata (e.g., public student GenAI datasets) to model (a) frequency of use, (b) learning-oriented versus shortcut-oriented use, and (c) perceived academic benefits and integrity concerns. Candidate methods include latent profile analysis to identify user typologies and multilevel regression to test how course- and institution-level supports relate to student outcomes. Implications highlight measurable leverage points for equitable GenAI pedagogy and governance.