ABSTRACT VIEW
Abstract NUM 953

AUTOMATIC COHERENCE ANALYSIS OF ORAL CONTRIBUTIONS IN ECONOMICS CLASSES: AN NLP-SUPPORTED FEEDBACK SYSTEM
A. Emter, C. Siegfried
University of Potsdam (GERMANY)
The central question of this study is how a Natural Language Processing (NLP)-based tool can support learners in vocational education in structuring their contributions in economics classes. The tool is based on the concept of self-regulated learning, the self-determination theory of motivation, and principles of critical-constructive didactics. The project utilizes a computer-supported feedback system that employs concept maps to promote the coherence of texts. Reflection on one's own contributions is considered a crucial factor in the development of subject competence such as economics competence, as learners are enabled to improve their arguments consciously. The promotion of argumentation is necessary in both vocational and general interpersonal contexts to ensure participation in society and profession and critical thinking ability.

Methodological Approach:
The project follows a mixed-methods approach. Verbal contributions are recorded, automatically transcribed, and analyzed using NLP technologies. Interactive concept maps visualize both the coherence of discussed topics and unexplored theme areas ("gray areas"). These "gray areas" were not considered in previous studies. The effectiveness of this feedback will be evaluated qualitatively (interviews) and quantitatively (standardized questionnaire).

Expected Results:
We expect that the visualization of arguments to discuss a topic will support learners in reflecting on their own contributions more consciously and structuring them more coherently. Additionally, the tool is expected to strengthen learner autonomy and independent work, ultimately leading to improved argumentation.

Relevant Implications:
The project contributes to the development of digital support systems in vocational education by linking NLP technologies with established didactic concepts. Such a system can support learners withing digital learning enviroments and support teachers in formative evaluation of argumentation.

Keywords: Natural Language Processing, Feedback, argumentation, economic competences, digital support.

Event: ICERI2025
Session: AI for Assessment and Feedback
Session time: Tuesday, 11th of November from 15:00 to 16:45
Session type: ORAL