Learning Analytics & Adaptive Learning

Duration: January 2022 – December 2025
Status: Completed
Educational Level: Tertiary Level
Topic: Artificial Intelligence AI, Digital Tools
Keywords: Digitised Learning Modules

Initital Situation

Individualised learning and teaching are a major challenge at the tertiary level, particularly in subjects such as methodology, where effective teaching requires a significant amount of effort. To address these challenges, the Institutes of Psychology and Sport Science (Faculty of Human Sciences) and the Institute of Mathematics (Faculty of Science) at the University of Bern have joined forces in this project. The aim is to make greater use of the technological possibilities of digital teaching and to develop innovative teaching and learning formats. The project focuses on the expansion and transfer of a Learning Analytics system that supports face-to-face teaching and provides students with helpful materials and personalised feedback to enable optimised, individualised learning.

Objectives

Through the Learning Analytics approach, many aspects of learning behavior can be measured digitally. However, the informational and predictive meaning of these data must be better understood to fully exploit their potential to support students in achieving learning goals. This BeLEARN project addressed the following questions:

1. How can large data volumes be processed efficiently and reduced to a manageable amount while still retaining relevant information? Can machine learning methods contribute to this?
2. Can data collected via Learning Analytics be conceptually linked to learning-relevant personality traits such as general cognitive ability, processing efficiency, effort, conscientiousness, or self-efficacy?
3. Which aspects of learning behaviour can be used to predict final grades in university courses?
4. Can individualised and personalised learning recommendations based on Learning Analytics data help students adapt their learning behavior during the semester?

Method

In the project, online assessments (initial, formative, and summative) were used that provided students with tasks related to lecture content and direct explanatory feedback on their answers. These online assessments enabled the automatic collection of performance data and processing times, as well as repeated surveys on learning behaviour, allowing the analysis of learning processes over entire semesters. The data were strictly pseudonymised: only students knew which pseudonym corresponded to their data, while instructors and researchers could not identify them. In addition, dashboards were developed to provide students with an overview of their performance and teachers with insights into learning progress.

Results

• Our research shows that Learning Analytics can effectively support individualised learning even in large groups.
• The online assessments made it possible to collect data-based information on learning behaviour and identify relevant personality traits such as conscientiousness or processing efficiency, improving the prediction of academic performance.
• Machine learning methods helped reduce complex data to a few meaningful, learning-relevant patterns, identifying behavioural types associated with different learning outcomes.
• The results enable personalised feedback that helps students adapt their learning behaviour early on, promotes self-regulated learning, and supports the achievement of learning goals.

Implemented Translation

To identify suitable partners for transferring our project into practice, discussions were held with various educational institutions and companies. The expectations and needs of learners, teachers, professionals in continuing education, and developers of teaching materials were continuously analysed. Based on these findings, the Swiss Learning Analytics Association (learning-analytics.ch) was founded. The association advises companies and educational institutions on evidence-based implementation of Learning Analytics and provides scalable, transferable open-source tools developed within these projects. The goal is to further develop and disseminate Learning Analytics in Switzerland in both research and application contexts. A first project with Swisscom is currently underway. In the long term, the association aims to establish itself as a platform for sustainable collaboration between research and practice in the field of Learning Analytics.

The project results have been continuously integrated into teaching practice, leading to a lasting anchoring of data-driven teaching development. Teachers use the insights to adapt teaching formats and support offerings. Students benefit from personalised feedback and targeted support — even in large groups. An extended dashboard helps students learn more effectively. The impact is measured through repeated evaluations, performance data, and feedback from both students and teachers.

Participating Institutions