Leibniz Research Initiative Digital Education

Formal teaching and learning processes in school and university science education are accompanied by digital media in various ways. Learning processes in the natural sciences are characterized by the construction of subject-specific ideas taking into account existing prior experience, the examination of natural phenomena in observations and experiments and the orientation towards competencies in the sense of subject-specific problem-solving skills. The hitherto largely untapped potential to support learning through the automated diagnosis of individual learning potential and learning processes and to accompany them in the sense of formative assessment opens up new opportunities to develop individualized, data-based formats for heterogeneous or inclusive learning groups and to research their effects. Against this background, the central objectives are to develop basic support options for learners and teachers and to develop new forms of digitally supported learning.

Concepts and concrete, exemplary implementations for the use of digital media are developed as part of this focus area. The focus here is on school settings on the one hand and university courses with comparatively large learning cohorts on the other. Due to current technological developments, concepts will be developed that use virtual reality and augmented reality technologies (VR and AR technologies), for example, and integrate motivating learning media (e.g. robots, embedded systems). Students on teacher training courses as well as general and vocational schools are involved in the individual activities in order to ensure the evaluation of the concepts and to use and discuss the research results directly with current and future teachers. In the area of collaborative learning platforms, concepts based on the LearnWeb infrastructure developed at L3S are being developed and evaluated, including in the context of teacher training at various universities and the Applied Machine Learning Academy in Hanover. Although the current focus is on subject-specific didactic research within electrical engineering and computer science, the feasibility of the concepts in other STEM fields within and outside the research initiative is particularly welcome.

The aim of this focus area is to investigate relevant factors (recommendations) that are either prerequisites or conducive factors for successful digitally supported learning processes and the further development of methods from data mining and learning analytics for the diagnosis and design of scientific learning processes. To this end, the extent to which the integration of certain digital methods can generate added value and risks for learning processes is being researched. In this way, analytical desiderata or sensible adaptations and further developments can be described, e.g. with regard to comparatively “small” learning groups or the quality of individual diagnoses. For example, the extent to which machine learning methods are helpful for the recommendation and structuring of media learning opportunities or whether the analysis of search behavior as part of learning processes (search as learning) can support them will be investigated. By analyzing the digitally generated data and supplementary learning success analyses, the information technology and didactic perspectives are included. This interdisciplinary approach makes it possible to identify the conditions for success and design requirements for educational programs in different settings.

In order to enable targeted research and implementation of data-supported teaching and learning, it is essential that the data collected is handled responsibly and ethically. Approaches such as “trusted learning analytics” (e.g. work by Prof. Dr. Hendrik Drachsler) lay an important foundation for the responsible and reliable use of learning analytics. This addresses, for example, aspects of transparent data handling and corresponding open communication as well as the involvement and consent of all persons involved. Investigating the convictions and ethical judgments of people in the various target groups and developing trust-building concepts based on these form the basis for wider use in different contexts and at different ages. Guaranteeing security and privacy plays a key role here. The aim is to develop transparent procedures for the analysis of learning behavior and the design of a data-driven concept for higher education didactics. This addresses both personal and structural conditions specific to individual departments and universities.