How Are University Students Tackling Collaborative Source-Based Writing in an Online Computer-Supported Collaborative Learning Environment? A Case Study

• In Event: Critical Aspects of Computer-Supported Collaborative Learning

Abstract

Objectives or purposes:
The objective of this study was to investigate the social interaction of university students while collaboratively writing source-based texts.

Perspective(s) or theoretical framework:
University students frequently engage in collaborative source-based writing tasks (e.g., literature overviews). Group interaction largely determines collaborative outcomes, but students often fail to attain productive interaction during collaborative tasks (Damşa, 2014; Gillies, 2019; Le et al., 2018).

Collaborating students should engage in high-quality epistemic and regulative actions. Epistemic actions concern the text’s content. Examples include ‘sharing information’ and ‘suggesting and discussing ideas for the text’. An example of a regulative action is ‘planning’. Eventually, students engage in ‘other’ actions, such as social chat (Damşa, 2014; Gillies, 2019).
Damşa (2014) showed that interaction patterns can be identified based on the degree to which groups focus on epistemic or regulative actions. Likewise, we will identify interaction patterns focusing on source-based writing tasks.
The research question guiding this study is: Which interaction patterns can be identified based on students’ interactions in collaborative source-based writing tasks?

Methods, techniques, or modes of inquiry & Data sources, evidence, objects, or materials:
Four groups of three university students in educational sciences participated. In 2021, data was collected during an academic writing course. In eight sessions, all groups synchronously wrote a literature overview using MS Teams to interact and Google Docs to write.
All sessions were recorded and transcribed including time stamps. Quantitative content analysis was applied. First, units of meaning (UoM) were demarcated. Next, two trained coders assigned one code (or ‘action’, such as ‘planning’) to each UoM (IRR: Cohen’s =.60, Percent Agreement=67.4 %). The coding scheme was based on Damşa (2014). In addition, process mining was used (Fluxicon, 2020). This enables visualization of actions and their sequences (see Figure 1). The quantitative data will be supplemented with video notes to determine the quality of each action.

Results and/or substantiated conclusions or warrants for arguments/point of view:
In this abstract, we only present the results for group 1’s first session because these are the only ones available so far. Figure 1 depicts a process map with paths connecting the actions that occurred during this session. The darker the action, the more UoM were coded as that action. The higher the frequency, the bolder the path. Frequencies are shown in the actions between square brackets and next to each path. The total amount of time (in minutes and seconds) spent on each action is shown between round brackets, along with the relative duration.
According to Figure 1, this group primarily engaged in social chat, mostly followed by ‘monitoring and status updates.’ The group also concentrated on less important aspects of the task (e.g., word count), on framing their text’s focus and on planning, also towards the end of their session. The occurring actions are to be expected at the beginning of a source-based writing task. Strikingly, the group spent more than half of their time chatting about not-task-related or less important aspects (resp. 47.21 % and 13,6 %), demonstrating poor time management. All further results will be included in the full paper.

Authors

• Bram De Wever, Ghent University
  Presenting Author

• Karen Putzeys, Ghent University
  Presenting Author