Article Preview
TopIntroduction
As a data-driven research area, learning analytics aims to observe students’ learning behaviors and intervene individual learning process through collecting and analyzing learning data (Brown, 2011). With the emergence of a variety of interactive learning environments in the last decade, learning analytics is being widely adopted in quantifying learning experiences, analyzing learning states, predicting learning effects and supporting personalized learning, etc. (Chatti, Dyckhoff, & Schroeder, 2012; Ferguson, 2012). Massive Open Online Course (MOOC) platforms like Coursera, Udacity, edX and Khan Academy have produced massive learning behavioral data, which offer great opportunities for applying learning analytics techniques to measure and predict learning performance (Chiou, & Shih, 2015; Moissa, Gasparini, & Kemczinski, 2015; Jiang, Williams, & Schenke, 2014). However, few studies on learning analytics are carried out to explore learning regularities in private learning environments (Fox, Patterson, & Ilson, 2014).
In recent years, many higher education institutions have joined Small Private Online Courses (SPOCs), which use MOOC materials to supplement classroom teaching, including high-quality video resources and rapid feedback (Fox, 2013). Forum is an important component of the interactions between teachers and students in SPOCs, which plays a critical role in improving students’ abilities of complex thinking and solving problems (Wen, Yang, & Rose, 2014). In order to promote the participation of students in collaborative learning, many teachers have incorporated students’ performance within discussions into grading rules to assess their comprehensive learning abilities (Romero, López, & Luna, 2013). Moreover, students’ opinions and feelings in forums are also quite valuable for teachers and researchers to understand their learning motivation and emotional states. In particular, positive emotions are beneficial for arousing students’ learning interests, and enable students to actively achieve learning objectives under the mental pleasure (Altrabsheh, Cocea, & Fallahkhair, 2015). Some relevant studies conclude that participation levels of discussion forums could indirectly contribute to students’ final academic achievement in MOOCs (Ezen-Can, Boyer, & Kellogg, 2015; Dowell et al., 2015).
Under the experimental context with massive discourse data, it is of great importance to investigate the relationship between discourses behavioral patterns (e.g., various interactive patterns and sentiment orientations) and academic achievements in SPOCs to reveal respective characteristics of different achieving groups. In order to promote the efficient collaborative learning in SPOCs, engagement patterns and sentiment states as well as their relationships with students’ academic achievements have become the critical research subjects, which will be analyzed and discussed in this study.