High-stake student drop-out prediction using hidden Markov models in fully asynchronous subscription-based MOOCs - Archive ouverte HAL Accéder directement au contenu
Article Dans Une Revue Technological Forecasting and Social Change Année : 2024

High-stake student drop-out prediction using hidden Markov models in fully asynchronous subscription-based MOOCs

Résumé

In this study, we analyze the learning behavior of 24,000 students in a fully asynchronous subscription-based MOOC platform using hidden Markov models (HMMs) to examine the relationship between learning motivation and student drop-out behavior. In contrast to previous findings, our results reveal that student drop-out is not necessarily correlated with low motivation, as students may drop out despite being highly motivated at the end of their learning journey. To design more effective student retention campaigns, educational decision-makers must consider the motivation level and target potential drop-outs with a low state of motivation. More specifically, our findings emphasize the need for early intervention to prevent students from dropping out, as it becomes challenging to stimulate motivation once it reaches its lowest state. By adopting our proposed methodology, decision-makers can gain a better understanding of the student drop-out process and make more informed student retention interventions.
Fichier non déposé

Dates et versions

hal-04542480 , version 1 (11-04-2024)

Identifiants

Citer

Dries Benoit, Wai Kit Tsang, Kristof Coussement, Annelies Raes. High-stake student drop-out prediction using hidden Markov models in fully asynchronous subscription-based MOOCs. Technological Forecasting and Social Change, 2024, 198, pp.123009. ⟨10.1016/j.techfore.2023.123009⟩. ⟨hal-04542480⟩
4 Consultations
0 Téléchargements

Altmetric

Partager

Gmail Facebook X LinkedIn More