Return
Computational thinking in collaborative programming discourse: an epistemic network analysis
DOI:10.1080/08993408.2026.2665188.png)
Abstract
En 中文
Collaborative programming offers advantages for fostering novice students’ computational thinking (CT) skills. Yet, there is a limited understanding of CT within a collaborative group and a lack of descriptive knowledge regarding CT in authentic learning processes through process-based approaches.
In this study, we applied the theoretical framework of collaborative problem-solving (CPS) to understand how CT emerged in group discourse. We investigated the association between CT skills and the social dimension of CPS skills, and its relation to the task performance.
The context of the study was a robotics programming workshop organised for 15–16-year-old students. The data included videos and log files which students’ activities were recorded. Through epistemic network analysis (ENA), we identified the co-occurrences of CT and CPS skills and compared the differences in discourse patterns between high- and low-performance groups.
We found that the high-performance groups discussed algorithm design and evaluation, while the discourse of the low-performance groups lacked examples of CT skills. Moreover, CT in two-way communication for building shared understanding was associated with the higher task performance.
Based on the results, we recommend that teachers design tasks and facilitate social interaction to encourage students to verbalise and practice CT skills within collaborative groups. For future studies, we suggest considering students’ background, analysing debugging processes where students build on outcomes of previous tasks to perform new ones, and examining how learning environments, such as robots, programming interfaces, and task design, influence the externalisation of CT skills in group discourse.
Keywords:
Computational thinking
collaborative programming
educational robotics
collaborative problem solving
K-12 education
epistemic network analysis

