Project description

Adaptive learning systems should support learners by, e.g., personalized feedback. In this dissertation project, a tool is developed that automatically analyzes learners' written answers based on Natural Language Processing methods. This tool will then be implemented in a learning environment that provides automatic feedback adapted to the learners' answers. Subsequently, the impact of this feedback on performance and user acceptance will be investigated.

Previous studies have repeatedly shown that learning success and the associated performance depend on learners’ ability to self-regulate learning: learners must continuously monitor their learning process (e.g., repeatedly evaluate if the applied learning strategy is best suited for the task at hand, and/or identify knowledge gaps) and adjust it if necessary.

However, various studies indicate that learners often have difficulties regulating their learning process independently. In these cases, feedback can be used as a supportive tool. However, feedback provided by humans is time-consuming, costly and not always objective – especially for tasks that require a textual response. Feedback generated by computer-based learning systems, on the other hand, is available quickly and at any time, is inexpensive and objective.  
Despite the advancing digitalization in education, there are so far only a few empirical studies investigating the influence of different types of feedback on learners' self-regulatory abilities and users' acceptance of this automatic feedback. As part of the "Human-Agent Interaction" network, this project contributes to closing this research gap.

COOPERATIONs

• Prof. Dr. Ulrike Padó, Stuttgart University of Applied Sciences
• Prof. Dr. Katharina Scheiter, Universität Potsdam