Dataset and Development of Learning Analytic Tool to Extract Manifestations of Students’ Agency from Texts of Comments from MOOCs

Abstract

The study is devoted to the automatic identification of manifestations of various components and sources of student agency from the texts of reviews of MOOCs, as well as descriptions of internal and external transformation among students in the process of studying MOOCs. To extract descriptions corresponding to individual, relational and contextual sources of students’ agency, a dataset of 3445 English-language comments on the most popular mathematics courses presented on the Udemy platform was generated, and additionally 1787 comments on practice-oriented MOOCs and entrepreneurship MOOCs were extracted to understand the descriptions corresponding manifestation of internal and external transformation in MOOC listeners. The paper proposes a methodological approach based on the use of natural language processing methods such as topic modeling, sentiment analysis and N-gram frequency analysis for extracting keywords and their combinations from MOOCs’ comments texts to describe the manifestation of the components of an individual source of student agency in the form of self-efficacy , increased sense of confidence in solving problems and motivation; components of the relational source in the form of support and accompaniment of the online course by the tutor with the help of quick answers and well-structured educational material; components of the contextual source in the form of the ability to make decisions when choosing alternative online courses, as well as descriptions of the manifestation of internal transformation of students, expressed in the transition from internal struggle - overcoming fears, uncertainty, difficulties in perceiving MOOC content to understanding the purpose of learning and external transformation, expressed in the texts of comments on MOOCs in the form of creating a new or changing the structure of an existing product, startup or business through a change in thinking.