Machine Learning in Predicting Undergraduate Student Attrition: A Scoping Review

• In Event: AERA Poster Session Thursday 7:45am
  In Poster Session: Poster Session: Statistical Theory, Quantitative Methods and Applications

Abstract

Undergraduate student attrition remains a persistent challenge in higher education, driven by complex, interconnected factors at the student, teacher, and institutional levels. Machine learning (ML) proves particularly effective for identifying at-risk students and predicting dropout patterns through analysis of complex, multifaceted data. This study uses a scoping review methodology to map ML applications for undergraduate student attrition prediction in the United States. We categorize studies by data sources (local, state, and national levels), ML models, and attrition outcome measures. Additionally, we present the mechanisms, strengths, and limitations of the ML models used in the literature. This review provides an overview of current ML approaches to student attrition prediction and identifies gaps for future research.

Authors

• Oluwaseun Peter Farotimi, University of Tennessee

• Oluwasayo Farotimi, Georgia Southern University

• Comfort Happiness Omonkhodion, University of Central Florida

• Funke Anuoluwapo Dada, Georgia Southern University

• Debbie L. Hahs-Vaughn, University of Central Florida

• Haiyan Bai, University of Central Florida