Supporting computer-based learning by investigating process data. Analysis of psychophysiological measurements and log file data.

Autor/in	Huber, Kerstin Sabine
Titel	Supporting computer-based learning by investigating process data. Analysis of psychophysiological measurements and log file data.
Quelle	München: Universitätsbibliothek der TU München (2023), vii, 131 S. PDF als Volltext kostenfreie Datei (1); PDF als Volltext kostenfreie Datei (2) Link als defekt meldenVerfügbarkeit Dissertation, Technische Universität München, 2023.
Sprache	englisch
Dokumenttyp	online; Monographie
URN	urn:nbn:de:bvb:91-diss-20230728-1706857-1-3
Schlagwörter	Lernen; Computerunterstützter Unterricht; Algorithmus; Dissertation; E-Learning
Abstract	The overall goal of the present research is to investigate the value of process data to understand learning processes and, thus, to promote computer-based learning. The findings of this thesis help educators identify learning processes in computer-based learning environments (CBLEs) aiming to provide meaningful, individualized support to learners (e.g., prompts; see chapters 2 and 7). The present dissertation includes two empirical studies investigating the manifestation of learning processes in process data. Both studies demonstrated how learning processes can be monitored and evaluated while learning with CBLEs. The goal of Study 1 was to validate psychophysiological measures to identify academic emotions and investigate their impact on learning outcomes. Electrodermal activity (EDA) and heart rate (HR) were measured during computer-based learning to monitor learners´ emotional states. Therefore, EDA, HR, and self-report data were gathered from 32 participants in a laboratory setting. To determine the manifestation of academic emotions in EDA, HR, and self-reports, negative emotions were induced using negative connotated learning materials about animal welfare. Participants reported their emotional states directly before and after learning, which were then collated with EDA and HR curves. A significant relationship was found between increased negative emotions and increased EDA and HR. Additionally, EDA turned out to be a significant indicator for learning performance. Furthermore, an explorative analysis revealed that boredom manifested in decreased HR. Study 2 investigated the impact of navigation behavior on learning performance using log file data and process mining analyses. Therefore, log files and self-report data were evaluated from 58 university students who used a CBLE for two weeks. The results showed a significant increase in learning with a very high effect size. A cluster analysis revealed two distinct learner groups, which differed significantly in their navigation behavior and learning outcomes. Here, the interactivity and the time spent on learningrelevant pages were meaningful indicators for learning outcomes, especially recall and transfer performance. In conclusion, the findings showed that beneficial and detrimental learning processes could be inferred from navigation behavior. Thus, the findings demonstrated that navigation behavior impacts learning outcomes. Moreover, Study 2 Abstract ½ iv presented a feasible approach to monitor and evaluate the interactivity and duration spent in a CBLE, which can be used to promote successful learning. In conclusion, this thesis demonstrates the importance of process data in investigating and supporting computer-based learning. The two empirical studies shed light on process data from different perspectives to obtain a comprehensive picture of learning processes. The findings contribute to identifying and evaluating learning processes in real time. Moreover, this dissertation presents important theoretical, methodological, and practical implications for further research and theory development. (übernommen).
Erfasst von	Deutsche Nationalbibliothek, Frankfurt am Main
Update	2024/1