Search form

B³ - Projects - AMIGO (Algorithmic Method for Intelligent Group Formation)


What is the AMIGO Project about?

Cooperative learning is an integral part of academic studies—almost every university seminar applies some sort of group work. This can be justified by two main arguments: First, the digitally connected world requires that social competences are trained at university to foster lifelong learning. Second, cooperative learning has been shown to be superior to individual learning with regard to learning success in countless empirical studies. However, one crucial aspect has been widely neglected in both research and practice: Which learners should be grouped together so that all learners benefit the most? In most instances, group formation is either performed randomly or through students’ choice, with both approaches being potentially problematic: Students tend to dislike random assignment, but self-determined group formation leads to mostly homogeneous groups—high-ability students stay among each other to form successful groups, while students, which lag behind, form groups that often do not achieve the course’s learning goals. Possible unintended outcomes include weak group performance, groups breaking apart, and unnecessary consultation workload for the teacher. Further, this approach also fosters exclusion of minorities or even bullying—thereby missing the enormous potential of diversity. An innovative approach is algorithmic intelligent group formation based on empirically evaluated criteria. Several personality traits (e.g. conscientiousness, extraversion), attitudes (e.g. motivation, team orientation), and skills (e.g. prior knowledge) are known to be influential for group performance and group satisfaction when optimally distributed. In our interdisciplinary research group, we developed the Moodle plugin ”group formation” that allows the diagnosis of relevant psychological criteria and formation of optimized groups. It is based on a non-linear optimization algorithm which respects homogeneous and heterogeneous criteria simultaneously while minimizing inter-group differences and increasing fairness. Empirical evaluation studies show that optimized groups performed better in weekly assignments and were more satisfied with the group work than randomized groups.

Mainz University and RWTH Aachen University have cooperated with Jacobs University to integrate intelligent learning group formation into Jacobs University´s innovative, interdisciplinary teaching and learning concept. The goal is to foster learning success and satisfaction for cooperative learning in higher education. The application of a Moodle plugin on group formation will be adapted to the circumstances and requirements at Jacobs University. We started analyzing cooperative learning scenarios at Jacobs University with respect to learning goals, task properties, as well as instructors’ and students’ experiences with the status quo and their respective needs. The results informed improvements of the plugin to optimally fit the needs at Jacobs University. The application of the plugin will be accompanied by design-based research methodology to advance our understanding of the optimal composition of learning groups and to rigorously evaluate the intervention effect of the plugin.

Technical advancements supported by the AMIGO project include:

  • Interdisciplinary groups: Instructors can define criteria that the algorithm has to consider (e.g., at least one student from subject A and one student from subject B in each group)
  • Choice of questionnaires: Instructors can choose from a set of predefined questionnaires that will be used for group formation
  • Minority protection: The algorithm makes sure that members of minorities are well represented in learning groups (e.g., at least two members of a minority in one group or none).
  • Language compatibility: Groups can be formed based on their preferred spoken language.

Instructors can manually reorganize groups in case of language barriers. This list of new features will be adjusted and extended to the needs of Jacobs University’s instructors.

After completion of this project, the tool will be integrated into the e-learning system at Jacobs University and will hence be ready for all instructors to use it in their courses.

Students who have expressed interest and approval of this method have made the following comments: the algorithm is ”more efficient” ; allows for ”quick group formation” and ”prevents stress in group formation.

Currently, we are conducting empirical studies in several seminars and lectures at Jacobs University, Mainz University, and RWTH Aachen University. The next steps will be data analysis and publication of the results. Further, we are working on the improvement of the plugin and preparing it for the roll-out at Jacobs University’s learning management system.

Project Chair: Henrik Bellhäuser (Johannes Gutenberg Universität Mainz)
Collaborators: Peter Baumann, Sonia Lippke and Stanislav Chankov (Jacobs University) and Rene Röpke (RWTH Aachen University).

  • Dr. Henrik Bellhäuser
    Johannes Gutenberg-Universität Mainz

  • Dr. Peter Baumann
    Jacobs University
  • Dr. Stanislav Chankov
    Jacobs University

  • Prof. Dr. Sonia Lippke
    Jacobs University

  • M.Sc. René Röpke
    RWTH Aachen University

AMIGO - Video with Dr. Bellhäuser

In this video, Dr. Henrik Bellhäuser talks about the B³ Project AMIGO (Algorithmic Method for Intelligent Group formation). The project explores the question of how to maximize cooperative learning among student groups. Learn about the algorithmic method for intelligent group formation that will be integrated into online learning platforms at Jacobs University this coming Fall 2021. This is the future of optimized cooperative learning!



Bellhäuser, H., Konert, J., Müller, A., & Röpke, R. (2018). Who is the Perfect Match? Effects of Algorithmic Learning Group Formation Using Personality Traits. Journal of Interactive Media (i-Com), 17(1), 65–77.

Konert, J., Bellhäuser, H., Röpke, R., Gallwas, E., & Zucik, A. (2016). MoodlePeers: Factors relevant in learning group formation for improved learning outcomes, satisfaction and commitment in E-learning scenarios using GroupAL. In K. Verbert, M. Sharples, & T. Klobucar (Eds.), Adaptive and Adaptable Learning: Proc. of the 11th European Conf. on Techn. Enhanced Learning (EC-TEL 2016) (pp. 390–396). Springer LNCS, Berlin.

Müller, A., Bellhäuser, H., Konert, J., & Röpke, R. (2021). Effects of group formation on student satisfaction and performance: A field experiment. Small Group Research, 104649642098859.