Using Topic Modeling for Institutional Assessment in Higher Education

• In Event: Innovations in Psychometrics Modelings, and Methodologies

Abstract

This study investigates the quality of the institutional assessment review process in a higher education institution, a process known as meta-assessment, by analyzing open-ended feedback from academic assessment reviews. We employed Latent Dirichlet Allocation (LDA), a topic modeling method, along with results from a rating scale to organize, analyze, and synthesize multiple years’ worth of academic program feedback. Our findings reveal underlying topics that offer valuable insights for incorporating feedback effectively within the assessment process while emphasizing continuous improvement. As institutional assessment systems evolve, our study highlights the importance of meta-assessment and topic modeling in enhancing assessment practices.

Author

• Lissette Tolentino, University of Florida