Title: Predicting student employment before graduation: a comparison of machine learning models

Authors: Linsey Hugo; William A. Young II; Marco Habermann; Ashley Y. Metcalf

Addresses: Ohio University, Athens, OH, USA ' Ohio University, Athens, OH, USA ' Ohio University, Athens, OH, USA ' Ohio University, Athens, OH, USA

Abstract: As universities are increasingly held accountable for students' career outcomes and competition for jobs increases, institutions need to understand which students are more likely to be employed upon graduation and why. This study aims to determine to what extent undergraduate student academic and experience employability signals - including major, GPA, co-curricular activities, and internships - can predict if a student secures full-time employment before graduation. Therefore, this study uses and compares the effectiveness of commonly recognised and advanced machine learning models, including logistic regression, discriminant analysis, decision trees, and neural networks. Results demonstrate that employment before graduation can be predicted with 74% accuracy with a neural network as the most accurate predictive model compared to the other approaches. Moreover, a sensitivity analysis identified co-curricular activities and majors as statistically significant variables predicting employment upon graduation.

Keywords: student employment; higher education; machine learning; logistic regression; discriminant analysis; decision trees; neural networks.

DOI: 10.1504/IJSSS.2024.144730

International Journal of Society Systems Science, 2024 Vol.15 No.2, pp.77 - 114

Received: 24 Jun 2023
Accepted: 25 May 2024
Published online: 28 Feb 2025 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article