During the last decades, collaboration in education and in the workplace has gained importance. At university, students are often prepared for this and are assigned collaborative tasks. Research showed that group interaction is largely determining the outcomes of collaboration (Damşa & Ludvigsen, 2016; Gillies, 2019; Qureshi et al., 2021). When students collaborate, they are expected to listen to their group members, clarify misconceptions, discuss each other’s ideas critically yet constructive, and help each other (Gillies, 2019; Rojas-Drummond & Mercer, 2003). Ideally, this interaction leads to strong group cohesion, and a general feeling of trust and confidence, which eventually results in success (Roseth et al., 2008).

Despite the fact that most university students are often familiar with working in small groups, writing texts collaboratively remains a challenging task. Le et al. (2018) noted that the majority of students do not know how to collaborate effectively. More in particular, they struggle with accepting opposing viewpoints, negotiating, providing elaborate explanations, and helping group members or accepting help. The quality of discourse in collaborative groups of students is often low: students do not activate prior knowledge, ask thought-provoking questions, nor elaborate on information spontaneously. However, this task-related discourse, concerning concepts and strategies is key for learning (Gillies, 2019; King, 2008). Furthermore, Le et al. (2018) state that effectively coordinating collaboration, including planning and time-management, is not self-evident for students. Also, in some groups, interaction is dominated by high status students, who are perceived as highly competent by their peers. These students see other, low-status group members as incompetent and passive. This leads to an imbalance in interaction between group members. Lastly, the authors noted that friendship between group members is an often overlooked, yet important determining factor for interaction. Being friends with their group members may cause less self-discipline and more off-task interaction (Le et al., 2018).

In this paper, students’ interaction during a collaborative writing task is examined. During collaborative writing, students’ interaction can consist of epistemic actions, such as sharing information and connecting ideas, or consist of regulative actions, such as planning and monitoring the writing process. A case study of Damşa and Ludvigsen (2016) shows that the degree to which groups of students focus on epistemic or regulative actions results in a specific interaction pattern. They distinguish an object-oriented, a discourse-oriented and an individual-action-based interaction pattern. The object-oriented pattern implies that students predominantly engage in epistemic actions, aimed at the collaboratively written text. The discourse-oriented pattern is reflected in more regulative and other, non-task related actions than epistemic actions. Here, students work toward shared understanding and elaboration on ideas, but fail to incorporate this developed knowledge into their text. In the individual-action-based pattern, students mainly engage in regulative actions, since they divide the labor by distributing reading and writing activities among the group members.

To conclude, students may have quite some collaborative experiences, but do not necessarily spontaneously interact effectively (Authors, 2020; Damşa & Ludvigsen, 2016; Gillies, 2019; Le et al., 2018). However, most research on collaboration focuses on the outcomes, while the interaction determining these outcomes remains underexposed. Therefore, the aim of the present study is to unveil the interactions in groups of university students who collaboratively write a text. In addition, the relationship between these interactions and the outcomes will be explored. The research questions (RQ) guiding this study are: (1) To which degree do university students engage in epistemic, regulative and other actions during a collaborative writing task?, (2) Which group profiles can be identified based on students’ interaction?, and (3) To which degree are these particular group profiles related to the quality of the collaboratively written texts?

Four groups of three master’s students in educational studies (mean age 20 years, SD = 0.54) participated in this study. As part of a course on academic writing, the groups wrote a literature overview, as if it were the theoretical framework for a study. All students received specific writing instruction lessons prior to the task. Students freely chose the topic of their paper within provided research domains. No resource articles were provided. The students wrote their text over the course of eight weeks in as many sessions as they preferred. Each session took place via MS Teams and Google Docs. Students’ interactions were recorded via MS Teams and transcribed using Subtitle Workshop 6.0b. The researcher will demarcate units of meaning (UoM) in the transcripts. In this study, a UoM is a set of consecutive expressions by one or more students aimed at the same goal. Each UoM was assigned one code, using a three-tiered hierarchical coding scheme based on Damşa and Ludvigsen (2016). At the highest level, it distinguishes between an epistemic, a regulative, and an other dimension. Each dimension contains categories of actions, which in turn comprises specific actions. Two independent trained coders will assign one code at the lowest level to each UoM and discuss their codes afterwards. Then, these codes will be aggregated to the highest level. The number of actions as well as the time invested in each action will be calculated. This will provide a picture of what the groups focused on primarily (e.g. ‘regulative actions’) in each writing session. Additional to this quantitative analysis, the researcher will perform an in-depth qualitative analysis through video notes to investigate the specific nature of the actions. In a previous study (Authors, 2020), a group of students invested relatively much time in ‘planning’. The video notes however showed that the group made a vague planning from the start and had to alter this planning multiple times throughout the process. This qualitative view adds to the understanding of how students interact. The combination of quantitative and qualitative data gives the opportunity to identify a group profile. The collaboratively written texts’ quality is assessed by two independent instructors. We will carefully attempt to relate the group profiles to the collaboratively written texts’ quality.

Since this is research in progress, not all findings can be reported at this stage. However, the intervention has taken place and all data has been successfully collected. We are currently in the middle of the coding process, so all results will be reported at the ECER conference. First, we will describe the degree to which students engage in epistemic, regulative and other actions (RQ 1) and identify a group profile (RQ 2). We will provide a clear overview of the number of epistemic, regulative and other actions, as well as the amount of time invested in these actions. Qualitative findings will be added to this quantitative picture. Based on these findings, we expect to identify different profiles for the groups. These profiles may be in line with the interaction patterns of Damşa and Ludvigsen (2016), as described above, or differ, as we used another task in our study. Finally, we will link these group profiles to the quality of the collaboratively written texts (RQ 3). Two groups wrote a low-quality text: group 1 scored 3.4 and group 2 scored 5.1 out of 10 points. Both group 3 and group 4 wrote high-quality texts and both scored 8.1 out of 10. It is not straightforward to formulate hypotheses about the proportion of epistemic and regulative actions and their relationship to the text quality. Low-quality texts can stem from problems in epistemic actions, such as excessively focusing on sharing information without linking concepts, or problems in regulative actions, such as an ill-defined plan of action, or other causes. However, this is merely an assumption which possibly will be confirmed or refuted based on our data.

Authors, 2020 Damşa, C. I., & Ludvigsen, S. (2016). Learning through interaction and co-construction of knowledge objects in teacher education. Learning, Culture and Social Interaction, 11, 1–18. https://doi.org/10.1016/j.lcsi.2016.03.001 Gillies, R. M. (2019). Promoting academically productive student dialogue during collaborative learning. International Journal of Educational Research, 97, 200–209. https://doi.org/10.1016/j.ijer.2017.07.014 King, A. (2008). Structuring Peer Interaction to Promote Higher-Order Thinking and Complex Learning in Cooperating Groups. In R. M. Gillies, A. F. Ashman, & J. Terwel (Eds.), The Teacher’s Role in Implementing Cooperative Learning in the Classroom. Computer-Supported Collaborative Learning. Springer. https://doi.org/https://doi.org/10.1007/978-0-387-70892-8_4 Le, H., Janssen, J., & Wubbels, T. (2018). Collaborative learning practices: teacher and student perceived obstacles to effective student collaboration. Cambridge Journal of Education, 48(1), 103–122. https://doi.org/10.1080/0305764X.2016.1259389 Qureshi, M. A., Khaskheli, A., Qureshi, J. A., Raza, S. A., & Yousufi, S. Q. (2021). Factors affecting students’ learning performance through collaborative learning and engagement. Interactive Learning Environments, 1–21. https://doi.org/10.1080/10494820.2021.1884886 Rojas-Drummond, S., & Mercer, N. (2003). Scaffolding the development of effective collaboration and learning. International Journal of Educational Research, 39(1–2), 99–111. https://doi.org/10.1016/S0883-0355(03)00075-2 Roseth, C. J., Johnson, D. W., & Johnson, R. T. (2008). Promoting early adolescents’ achievement and peer relationships: the effects of cooperative, competitive, and individualistic goal structures. Psychological Bulletin, 134(2), 223–246.