Adaptive Learning Mechanisms for Learning Management Systems: A Scoping Review and Practical Considerations

TL;DR

This paper systematically reviews adaptive learning mechanisms in LMSs, highlighting their integration methods, limitations, and proposing future directions for system-independent specifications and user-friendly authoring tools.

Contribution

It provides a comprehensive review of adaptive learning approaches, tests hypotheses about their data usage and customization, and suggests new research directions for system-independent models.

Findings

Adaptive mechanisms rarely consider existing data.

They support limited data processing mechanisms.

Users are rarely able to customize how mechanisms work.

Abstract

Background: Traditional Learning Management Systems (LMS) usually offer a one-size-fits-all solution that cannot be customized to meet specific learner needs. To address this issue, adaptive learning mechanisms are integrated either by LMS-specific approaches into individual LMSs or by system-independent mechanisms into various existing LMSs to increase reusability. Objective: We conducted a systematic review of the literature addressing the following research questions. How are adaptive learning mechanisms integrated into LMSs system-independently? How are they provided, how are they specified, and on which database do they operate? A priori, we proposed three hypotheses. First, the focused adaptive learning mechanisms, rarely consider existing data. Second, they usually support a limited number of data processing mechanisms. Third, the users intended to provide them, are rarely given the ability to adapt how they work. Furthermore, to investigate the differences between system-independent and LMS-specific approaches, we also included the latter. Design: We used Scopus, Web of Science and Google Scholar for gray literature to identify 3370 papers published between 2003 and 2023 for screening, and conducted a snowball search. Results: We identified 61 relevant approaches and extracted eight variables for them through in-depth reading. The results support the proposed hypotheses. Conclusion: Based on the challenges raised by the proposed hypotheses with regard to the relevant user groups, we defined two future research directions - developing a conceptual model for the system-independent specification of adaptive learning mechanisms and a corresponding architecture for the provision, and supporting the authoring of these mechanisms by users with low technical expertise.