Modeling Achievement Gaps in Higher Education
Abstract: In this talk, I will present a data analysis of achievement gaps in higher education using logistic mixed-effects models on a large longitudinal dataset of student records from 2016 to 2023. I account for within-student variance using random effects and model nonlinear effects of high school GPA. Model performance is compared with traditional logistic regression and other predictive models for both new and previously observed students. I also compute predicted probabilities and corresponding risk percentiles to examine disparities across race, sex, and first-generation status and demonstrate how this modeling strategy can support further institutional research.
Biography: Rick Brown is an assistant professor of mathematics and the director of the data science program at Southern Utah University. He earned a B.S. in Mathematics from California State University, Chico, an M.S. in Statistics from Colorado State University, a PhD in Applied Mathematics from the University of Montana, and worked as a Mathematical Statistician for the U.S. Census Bureau between graduate school stints. Rick has research interests in applied statistics, Bayesian methods, and inverse problems and spends time mentoring undergraduate students with their statistics and data science research projects.