Epistemic Network Analysis and Topic Modeling for Chat Data from Collaborative Learning Environment [Konferenzbericht] Paper presented at the International Conference on Educational Data Mining (10th, Wuhan, China, Jun 25-28, 2017).

Autor/inn/en	Cai, Zhiqiang; Pennebaker, James W.; Eagan, Brendan; Shaffer, David W.; Dowell, Nia M.; Graesser, Arthur C.
Titel	Epistemic Network Analysis and Topic Modeling for Chat Data from Collaborative Learning Environment [Konferenzbericht] Paper presented at the International Conference on Educational Data Mining (10th, Wuhan, China, Jun 25-28, 2017).
Quelle	(2017), (8 Seiten) PDF als Volltext kostenfreie Datei Verfügbarkeit
Zusatzinformation	Weitere Informationen
Sprache	englisch
Dokumenttyp	gedruckt; online; Monographie
Schlagwörter	Network Analysis; Computer Mediated Communication; Cooperative Learning; Scores; Prediction; Data Analysis; Epistemology; Learning Processes; Academic Achievement; Introductory Courses; Psychology; Undergraduate Students; Personality Problems; Tests; Achievement Gains; Course Descriptions + Suchen Sie Ihr Suchwort? Netzplantechnik; Computerkonferenz; Kooperatives Lernen; Vorhersage; Auswertung; Erkenntnistheorie; Learning process; Lernprozess; Schulleistung; Einführungskurs; Psychologie; Examination; Prüfung; Examen; Achievement gain; Leistungssteigerung; Kursstrukturplan
Abstract	This study investigates a possible way to analyze chat data from collaborative learning environments using epistemic network analysis and topic modeling. A 300-topic general topic model built from TASA (Touchstone Applied Science Associates) corpus was used in this study. 300 topic scores for each of the 15,670 utterances in our chat data were computed. Seven relevant topics were selected based on the total document scores. While the aggregated topic scores had some power in predicting students' learning, using epistemic network analysis enables assessing the data from a different angle. The results showed that the topic score based epistemic networks between low gain students and high gain students were significantly different (t = 2.00). Overall, the results suggest these two analytical approaches provide complementary information and afford new insights into the processes related to successful collaborative interactions. [This paper was published in: X. Hu, T. Barnes, A. Hershkovitz, L. Paquette (Eds), "Proceedings of the 10th International Conference on Educational Data Mining" (pp. 104-111). Wuhan, China: EDM Society.] (As Provided).
Erfasst von	ERIC (Education Resources Information Center), Washington, DC
Update	2020/1/01