Research Analyst II Ohio University Ithaca, New York, United States
Session Abstract: Most retention studies focus on a binary outcome—retain or not—over a limited timeframe. This binary ignores two critical pathways: transfer and early graduation. This study uses multinomial logistic regression in R to model undergraduates'' probabilities of enrolling, not enrolling, transferring, or graduating across six academic years at a large public four-year university. Predictors include student background (ACT/SAT scores, high school GPA, gender, race/ethnicity, first-generation), financial factors (unmet need, cost of attendance, Pell eligibility, etc.), academic progress (course drops, withdrawals, failures, and course non-success rate), and census variables. Findings can inform institutional collaboration (e.g., academic support, residential life, budget, etc.) to identify at-risk students for early interventions, anticipate budgetary needs, and target high-risk courses. This study contributes methodological innovation and practical strategies for student success.
