Understanding student teachers’ collaborative problem solving: Insights from an epistemic network analysis (ENA)

Highlights

• Epistemic network analysis was used to analyze students' collaborative problem solving processes.

• Both the high and low academic performance groups worked to maintain positive communication.

• Differences in CPS strategies led to the differences between the high and low academic performance groups.

Abstract

Collaborative problem solving, as a key competency in the 21st century, includes both social and cognitive processes with interactive, interdependent, and periodic characteristics, so it is difficult to analyze collaborative problem solving by traditional coding and counting methods. There is a need for a new analysis approach that can capture the temporal and dynamic process of collaborative problem solving in diversity online collaborative learning context to provide some insights into online collaborative learning design. During an eight-week semester, a total of 42 student teachers participated in two online collaborative learning activities. Student teachers' discourse data were collected, and the data were coded based on a collaborative problem solving assessment model. This study used Epistemic Network Analysis (ENA) to explore the collaborative problem solving processes of student teachers in different online collaborative learning tasks. The results showed that both the high and low academic performance groups worked to maintain positive communication, but the students in the high academic performance groups negotiated on ideas while the students in the low academic performance groups focused on sharing resources/ideas. Moreover, fine-grained centroid analysis on a weekly basis showed that the high academic performance groups began by maintaining positive communication, and ended by negotiating ideas, while the low academic performance groups began by sharing resources/ideas and ended by regulating problem solving activities. Finally, the implications, limitations, and future research were discussed.

Introduction

In the last 20 years, technological, economic, and social development have put new demands on peoples' key competencies, and collaborative problem solving (CPS) has been recognized as an essential component (e.g., von Davier, Hao, Liu, & Kyllonen, 2017). When learners are faced with complex tasks, they usually rely on the strength of team instead of individual ability to fulfill tasks collaboratively. Although the differences among individual students and the diversity of learning scenarios complicate online collaborative learning, it has been widely adopted. In collaborative learning settings, group members can effectively divide their work, increase knowledge and experience in the process of mutual consultation, elaboration and discussion, generate creative ideas, and thus reach a common understanding and successfully complete tasks (e.g., Damşa, 2014). At the same time, social interaction can help individuals develop cognitive and problem solving skills (e.g., Vygotsky, 1986; Wittrock, 1989).

Organization for Economic Cooperation and Development (OECD) has defined CPS as “both a cognitive and social process” (OECD, 2013, p. 6) and thus consists of two broader skill sets: social and cognitive skills (e.g., Liu, Hao, von Davier, Kyllonen, & Zapata-Rivera, 2015). CPS competency is conducive to learners' deep learning and effective problem solving, and successful collaboration depends on effective interaction between team members, as well as the competency of team members (Griffin, McGaw, & Care, 2012). What has not been made clear is how CPS competency is composed and how cognitive and social skills interact while at the same time influencing subsequent actions. Therefore, more research is needed on the process and key elements of CPS in multi-stage of online collaborative learning.

Meanwhile, since CPS is a socio-cognitive, interactive, interdependent, and temporal process, it is challenging to develop a standardized assessment of it (e.g., Swiecki, Ruis, Farrell, & Shaffer, 2020; von Davier et al., 2017). Questionnaire and self-report have traditionally provided limited understanding for CPS. Coding and counting can give insight into individual contributions. Computer supported collaborative learning (CSCL) takes place in a digital environment and students usually use text, language and video to communicate without having to face each other, the learners' traces and discourse data can be recorded. By capturing the trace data of learners, the process of collaborative problem solving can be analyzed in depth (e.g., von Davier et al., 2017). Using educational data mining and learning analytics to investigate collaborative problem solving process can detect and predict the quality of teamwork (e.g., Scoular & Care, 2020). Epistemic Network Analysis (ENA), as a technique of learning analytics, enables both quantitative and qualitative analysis of complex collaborative interaction process (e.g., Shaffer, 2017; Siebert-Evenstone et al., 2017). ENA is a powerful tool for evaluating CPS because of its capacity to reveal interactivity, interdependence and temporal process in collaboration (e.g., Swiecki et al., 2020).

It is also challenging to investigate collaborative problem solving in teacher education. The important role of collaboration in teacher learning and professional development has been recognized in many studies (e.g., Zhang, Liu, Chen, Wang, & Huang, 2017; Schuster, Hartmann, & Kolleck, 2021). Collaborative learning is also a commonly used teaching method in teacher education courses, as it is assumed to support student teachers in acquiring practical knowledge and skills needed for their future professional life (e.g., Willegems, Consuegra, Struyven, & Engels, 2018). How teachers perform collaborative problem solving in collaborative learning context has been investigated. For example, Kwon, Song, Sari, and Khikmatillaeva (2019) had studied pre-service teachers' collaborative problem solving process and their effects on solution quality in a computer education course. They conducted content analysis approach to investigate pre-service teachers' collaborative problem solving behaviors and calculated the Pearson coefficients between the frequency of behaviors and the solution scores to reveal the relationship between the collaborative problem solving process and the quality of the solutions. Griffin, Jones, and Kilgore (2006) used a qualitative approach to reveal pre-service teachers' perceptions of a collaborative problem solving activity. Recent studies on teachers' collaborative problem solving have paid attention to the interactivity, interdependency, and temporality of collaborative actions. For example, Ouyang, Hu, Zhang, Guo, and Yang (2021) used a multi-method approach combining quantitative content analysis, lag-sequential analysis, and frequent-sequence mining to quantitatively measure the interaction processes of teachers in collaborative problem solving activities. Combining social network analysis, content analysis and lag sequential analysis, Zhang et al. (2017) explored primary school teachers' interactive networks and social knowledge construction behavioral patterns in online collaborative learning activities. In line with the increasing emphasis on complex collaborative process analysis in teacher collaborative learning (Walkoe & Luna, 2020), more research is needed to explore student teachers’ fine-grained CPS process in order to achieve high collaborative performance.

Given the need for a fine-grained analysis about CPS process in a collaborative learning environment, this study aims to leverage the affordances of ENA to examine student teachers’ CPS process and identify the temporal, interdependent, and cyclical patterns of CPS among different academic performance groups. This study sheds light on the activity design and the application of learning analytics in CSCL settings.