Repeated Peer Assessment in Long-Term Group Work: Uncovering Student Views for Improved Implementation

Research Paper Recommended citation: Yström, A., Särner, E., & Axin, M. (2025). Repeated Peer Assessment in Long-Term Group Work: Uncovering Student Views for Improved Implementation. In Kangaslampi, R., Langie, G., Järvinen, H.-M., & Nagy, B. (Eds.), SEFI 53rd Annual Conference. European Society for Engineering Education (SEFI), Tampere, Finland. DOI: 10.5281/zenodo.17631830. This Conference Paper is brought to you for open access by the 53rd Annual Conference of the European Society for Engineering Education (SEFI) at Tampere University in Tampere, Finland. This work is licensed under a Creative Commons Attribution-NonCommercial-Share Alike 4.0 International License. REPEATED PEER ASSESSMENT IN LONG-TERM GROUP WORK: UNCOVERING STUDENT VIEWS FOR IMPROVED IMPLEMENTATION A Yströma, 1 , E Särnerb, M Axinc a Linköping University, Linköping, Sweden, 0000-0003-4344-1685 b Linköping University, Linköping, Sweden c Linköping University, Linköping, Sweden Conference Key Areas: Engineering skills, professional skills, and transversal skills/ Digital tools and AI in engineering education Keywords: peer assessment, group work, group dynamics ABSTRACT This study investigates students' perceptions of repeated peer assessments evaluating transversal skills in a group project setting. We draw on the introduction of a digital tool to monitor group dynamics in semester-long projects for 3rd year students in a Swedish university context implemented in 2022-2024. We discuss the student’s perception of peer assessment, and critical implementation challenges. 1 INTRODUCTION Effective assessment of group work requires close, continuous observation of student behaviours. Repeated peer assessments may provide insights into group dynamics that examiners otherwise would not have access to (Conway et al., 1993), and help in uncovering problematic group dynamics such as “freeriders” or “bad apples” (Felps et al., 2006; Sridharan et al., 2018), although it should not be considered a universal panacea to such problems (Topping, 1998). Although academics are generally positive towards the potential insights that can be gained from such assessments, widespread implementation is often hindered by a lack of time, knowledge, or interest among teachers, as well as challenges in obtaining relevant insights from students at the right time, and concerns about the honesty and sincerity of student feedback (Adachi et al., 2018; Fleckney et al., 2025). 1 Corresponding Author A Yström [email protected] In the Swedish context, the use of peer assessment of “transversal skills” (Berdanier, 2022) in group work in higher engineering education is increasing, but at the same time challenging traditions in the Swedish teaching culture concerning the inclusion of students in evaluation and assessment of learning outcomes (Adachi et al., 2018). Some critical insights from previous research on peer assessment relates to how the feedback and assessments made where in fact utilized and acted upon, either by teachers or the students (Adachi et al., 2018), drawing attention to the need for clarity about the purpose and processes of assessment (Sridharan et al., 2023). Otherwise, students may question the usefulness of the assessment and the efforts it requires in relation to what is potentially gained (Hanrahan & Isaacs, 2001). In addition, students’ literacy in giving and taking feedback has been raised as a challenge (Adachi et al., 2018; Machera, 2017), indicating that training and scaffolding is required to support the students in the process, dependent on the form and purpose of the assessment (Carless & and Boud, 2018; Sridharan et al., 2023; Topping, 1998). Increasingly, studies are acknowledging the multitude of factors that contribute to successful implementation of peer assessment (Ashenafi, 2017; Fleckney et al., 2025; Sridharan et al., 2023), and specifically exploring the students’ perceptions of the assessment, however, mainly through quantitative study designs (e.g., Sridharan & Boud, 2019). Further understanding students’ perceptions and motivations to engage in selfand peer assessment has been called for as an important area of further research to develop better implementation strategies (Adachi et al., 2018). Addressing this call, this paper centers on the research question: How do students perceive repeated peer assessments in long-term group projects, and how do such perceptions relate to implementation objectives? We draw on a 3-year study of the implementation of a digital tool supporting peer assessment of individual contributions in long-term group projects at an engineering program at a Swedish university. The repeated peer assessment enabled us to monitor the development of group dynamics and identify major deviations that might warrant teacher intervention, and the assessment purposively did not have an impact on grading. The introduction of a peer assessment tool was motivated by experiences from previous years where problematic group dynamics and uneven contributions of group members had been uncovered at very late stages of the projects, when it was too late for interventions. To that end, the repeated assessments provided a “safe” channel for students to raise concerns about problematic group dynamics, as the feedback from the assessments were only visible to the teachers, and the purpose was thus not to provide student-student feedback. Similar to the study by Hanrahan and Isaacs (2001) and more recently Grammenos et al. (2019) and Sridharan et al. (2018), we were mainly interested in the students’ perception and reflections related to peer assessment, specifically related to implementation. 2 METHODOLOGY 2.1 Study design We conducted a case study (Bryman & Bell, 2007; Creswell et al., 2003) of the introduction of a digital tool enabling peer assessment of transversal skills and individual contributions in project groups at a Swedish engineering program. The study was carried out in 2022-2024, covering all stages from tool/survey development until implementation. This type of study design allowed us both qualitative and quantitative insights related to students’ perception of peer assessment over the years. Data collection Qualitative data on student perceptions was collected primarily through the openended questions included in the survey. To triangulate this data (Fusch et al., 2018), we also drew upon information gathered during focus group interviews with project managers which gave a more informal understanding of the “status” in the group at the time from their perspectives, as well as informal conversations in the teaching team (see table 1 for details), where additional interactions with students and project stakeholders were discussed and notes were taken, which resulted in a more fullfledged understanding of students’ perceptions. Table 1. Data collection Phase Data source Tool/ survey development (2022) Trial run (2023) Refined implementation (2024) Survey/Digital tool No of groups: 2 1 assessment point (after Gate 4/project completion) Qualitative (open comments) and quantitative data (peer assessments) No of groups: 12 4 assessment points (Gates 1-4) Qualitative (open comments) and quantitative data (peer assessments) No of groups: 10 3 assessment points (Gates 2-4) Qualitative (open comments) and quantitative data (peer assessments) Focus group interviews with project managers (students) Informal conversations with 2 project managers (notes taken) 4 interviews (12 project leaders) in conjunction with assessment points (notes taken) 4 interviews (10 project leaders) in conjunction with assessment points, also after project completion (notes taken) Conversations in the teaching team (3 teachers) Sporadic discussions around survey development, in particular itemand scale construction (notes taken) Recurring discussions before and during the course, specifically in conjunction with assessment points about student perceptions of the assessments (notes taken) Recurring discussions before and during the course, specifically in conjunction with assessment points about student perceptions of the assessments (notes taken) Survey/Digital tool Aside from the quantitative data from the survey, in this study we center on the qualitative data extracted from the open-ended comments provided on survey items, as well as two open-ended questions at the end of the survey (Q1: Any comments with regards to the assessments made? and Q2: Comments, suggestions for improvements and other opinions that can be useful to improve the assessment tool). Focus group interviews The focus group interviews complemented the data from the survey by providing perspectives from the project managers (who were also students) on group dynamics, typically conducted face-to-face. These informal forums encouraged project managers to share current project group happenings, discuss best practices, and seek advice on common issues. One author guided the discussions with questions like, “How is your group doing right now?” and “Are you facing any particular challenges in your groups?” Informal conversations Initially, the teaching team focused on developing the survey and digital tool, including item creation and scale discussions. As the study progressed, conversations naturally shifted to analyzing survey data at different assessment points, discussing the comments provided by the students, identifying patterns in group dynamics, and determining which groups might need interventions or should be monitored until the next assessment. Data analysis To gain a comprehensive understanding of the students’ perspectives on the peer assessment, recurrent patterns and themes (Strauss & Corbin, 2008) were inductively identified in the open comments from the survey, looking across the years of development and implementation. This qualitative data was then triangulated with insights from the focus group interviews and informal conversations within the teaching team. By cross-referencing these sources, we were able to grasp student perceptions, trace developments in group dynamics, and identify common challenges related to the assessment, ensuring trustworthiness in our interpretations (Lincoln & Guba, 1985). 2.2 The peer assessments The assessments were conducted in third-year group projects of a design and product development engineering program at a Swedish university, involving approximately 90 students each year. Students worked in groups of 6-8 to develop new product or service concepts, without choosing their group members. Many projects involved external stakeholders, requiring more professional project management efforts to keep the projects on track. Over the 20-week duration, groups typically distribute specialized tasks to smaller sub-groups, while still requiring continuous involvement in overall project activities. Each group was required to select a project manager among the group members. The assessment was introduced to the students at the first day of the course, outlining the general aim of the assessment to help the teachers uncover problematic group dynamics in due time to make interventions, and the development of the tool being part of a pedagogical research project. To further increase the sense of privacy for each student, and encouraging honest responses, it was clearly communicated that the outcomes of the assessments were only made available to the teaching team, not the project groups themselves, unless something was considered so problematic that it warranted an intervention from the teachers. Thus, it was made clear that the assessment made no impact on the grading, rather it was incorporated as part of the compulsory deliverables within the course to ensure that all students filled out the assessments at each occasion. The digital link to the assessment was sent out to each team with approximately a 1week window to complete the assessment. The assessment could be completed at any time, alone or together, during this period. The survey items used in the peer assessment were adapted from Lejk and Wyvill (2001) and Conway et al. (1993), modified to suit our context. Students ranked themselves and their group members on a 3-graded scale (not at all (0)/not well (1)/well (2)/very well (3)) for each item). On average, it took between 5-15 minutes to complete the assessment, taking longer in on the first occasion compared to the final one. On occasion, reminders were sent on the final day to groups where individuals had not yet completed the assessment. An example of how the quantitative data was visualized can be found in Figure 1, where the assessments (including self-assessments) in the group were compiled in radar diagrams showing the averages for each individual on each item. Such illustrations supported identification of major variations at the 3-4 assessment occasions during the project life span. Assessment occasion 1 Assessment occasion 2 Assessment occasion 3 Explanation of the radar diagrams: Each individual is represented by one coloured line, and as an example in this particular case shows one individual (student 7, dark blue line more towards the center) consistently scoring lower than the rest of the group in all the three assessments. It is however noteworthy that the general individual scores start relatively high at Occasion 1, diminish by Occasion 2, to become slightly more distributed at Occasion 3, where Occasion 2 typically occurring around 4 weeks into the course, could be a reflection on the group entering the storming phase according to Tuckman’s group development model. Fig. 1. Example of visualization of variations in group dynamics, based on peer assessments (group from 2024) 3 FINDINGS: STUDENT PERCEPTIONS ABOUT THE PEER ASSESSMENT Similar to findings from previous research (e.g., Hanrahan & Isaacs, 2001), we could see that the assessments appeared to trigger increased self-reflexivity and selfevaluation, spurred by the fact that the students were evaluating themselves and the other students, as exemplified by this student: ”It [the assessment] gives a possibility to reflect about how the group is doing and how oneself is contributing.” To dig deeper into those kinds of reflections, as visible in our data, our analysis has centered outlining four central points that the students’ raised related to the format, execution and scaffolding of the peer assessments, which merits further consideration in future implementation of peer assessments. 1) Initial skepticism and personal discomfort Many students experienced some initial skepticism and awkwardness towards evaluating their group members. Some groups had even agreed on not reporting problems and to give everyone the same score in the assessment, which would then be counterproductive for the purpose of survey. This student was of the opinion that problems should first be discussed with the project leader and solved within the group: “The survey feels unnecessary personal and contributes to thinking about something that might not be a problem. But it is good and clear to know where to turn if you are discontent and that it is clear what is expected by everyone. Unfortunately, we agree that problems should primarily be raised with the project leader and solved within the group before the steering committee is informed. That has worked well!” 2) Uncertainty in utilizing assessment outcomes The open comments from the survey also showed that the opportunity to evaluate the individual contributions from each group member was appreciated and perceived to be a potentially useful if the group could not function properly. Still, questions remained about how the assessment itself, as well as the results, could in fact be useful: ”Yes, it [the survey] is good, but it is unclear how you act if you see that something is not working within the group.” Especially the groups that experienced problems found it unclear as how to handle the problems that were reported as stated by one student: “I am of the opinion that it [the survey] hasn’t had any effect. Right now we experience conflicts in the group because of unequal workloads and despite discussions it has not improved.” 3) Need for dialogue to resolve group dynamics Some students seemed disappointed that the assessment itself did not resolve the conflicts or irritations within the groups, indicating that the purpose of the assessment could have been clearer. Frustrated with the limitations of the assessment tool itself to provide solutions, students called for complementary ways of addressing group dynamics: “I actually think it would be better to discuss in the group. If someone is perceived to misbehave or not perform as agreed, it is better to bring it up with the person in question so that you have a chance to change than to fill out a survey. I think that the atmosphere in the group is better if you have an open dialog.” Another student indicated the need for dialogue with stakeholders also outside of the project group as an important way to move forward: ”I think there should be a complementary meeting with you in the steering committee maybe once every two weeks with each project leader where the project leader can explain around the group dynamics. Sometimes it is hard to give a grade on group dynamics.” 4) Timing of and frequency of assessment occasions In the first year of full implementation (2023), we had four assessment occasions following the project management gate structure integrated in the course design, but the time between the third and fourth assessment was perceived to be too long. In this period, the project work became more intense, and more conflicts are likely to emerge or intensify under such circumstances, which was also recognized by some of the students: “It is quite a long time between gate 3 and gate 4 and I know that there are groups where conflicts and such have arisen during this time. It had been appropriate to have an evaluation halfway between gate 3 and gate 4 to have the possibility to act before the project is over.” Clearly, the timing of the assessment occasions matter and requires careful consideration. 4 DISCUSSION AND CONCLUSIONS This study contributes to engineering education research and practice by examining students' perceptions of peer assessment in group work in a Swedish context, particularly in identifying problematic group dynamics. As such it contributes to an increasing body of research specifically exploring student perceptions of selfand peer assessment, beyond performance metrics (Grammenos et al., 2019; Hanrahan & Isaacs, 2001; Sridharan et al., 2018), where few studies have adopted a qualitative methodology that could enable a more empathetic and effective implementation. Ultimately, the teaching team found that the assessments provided valuable insights into the groups that would otherwise have remained undisclosed (Adachi et al., 2018), as we could now make note of perceptions of individual team members’ contributions at multiple occasions during the project, as well as tap into on-going reflections on what is going on in the groups through the open ended comments in the survey. Based on the analysis of student perceptions, and the underlying implementation objectives in our case, further experimentation with these kinds of assessments and tools is warranted despite the extra time and effort it requires. This statement should be considered in light of the Swedish assessment culture, where formative assessments (promoting learning through continuous feedback) is the dominating approach, and yet, students report a limited influence on assessments (as compared to e.g. UK and Australia) (Pereira et al., 2017; Simper et al., 2022). While no major interventions had been organized as a direct result of the peer assessments (no group’s assessment uncovered dynamics that warranted direct intervention), the assessments still contributed to more informed group discussions led by the teaching team, as well as other project stakeholders at each gate (milestone) during the project. It is also clear that while repeated selfand peer assessments as implemented in our study may allow teachers to track of group dynamics, as pointed out by the students, these assessments only help identify issues but do not provide solutions, as noted also by e.g. Topping (1998) and Sridharan et al. (2023). To resolve potential discrepancies between student perceptions and implementation objectives, we conclude that implementation requires thoughtful adjustments to course design and components for a more comprehensive impact (Fleckney et al., 2025; Sridharan et al., 2023), and is contingent on the specific purpose and process of peer assessment (Topping, 2010), and the specific context (. We outline three critical areas for teacher consideration to ensure better alignment between student perceptions and implementation objectives: 1) Communication and framing - implementation of such an assessment tool requires clear communication about its purpose and guidance on handling any issues that arise or conflicts that emerge, not suggesting that the assessment itself is a cure-all. Building on e.g., Sridharan et al. (2023), we note that effective communication about the assessment is crucial to maximize its impact, as well as how to best prepare the students to do the assessment, which may require additional scaffolding to ensure high-quality input from the students. Additionally, clarifying and discussing the assessment's purpose with the students on multiple occasions, not just once, can support students in becoming more comfortable with the task. Explicating how the teachers may use the information to intervene in group dynamics is also important, indicating that it is not only a paper construct but can result in actions. The feedback provided to students based on the assessments should be aligned with the specific implementation objectives in each case. 2) Timing of the assessments – it is essential for the teachers to carefully consider the timing of the assessment in relation to the group task at hand, to be able to pick up on potential conflicts early on, as well as to monitor and support the group in handling them in due time. The issue of timing of peer assessment has previously been noted by e.g., Brooks and Ammons (2003), where repeated assessments appear to be useful in addressing free-riding problems in groups. Here, given our focus on group dynamics in the peer assessments, our study highlights the usefulness of multiple assessments early on in the group development process, however, this must also be adapted to course design and specific group task characteristics. 3) Utilization of the assessment outcomesWhile the digital tool was appreciated as a channel to air potential problems, it was not supportive in resolving problems when they occurred. This concern is not unique to our study, and can be noted also in studies where e.g., the outcomes of the assessments are communicated not only to the teachers but also to the students (see e.g., Sridharan et al., 2023) To address this, additional course components could be introduced to more fully engage with the outcomes of each assessment, sharing the outcomes directly or indirectly to the groups for further processing. Doing so could support further discussions within the groups (or at least triggered an open discussion about the state of the group). Introducing additional course elements on e.g., conflict management and group development could provide in-class opportunities for groups to discuss current and desired group dynamics in conjunction with the recurring assessments, and it would require additional scaffolding and training to develop students’ feedback literacy, as noted also by e.g., Panadero et al. (2016) and Hoo et al. (2022). On a final note, we acknowledge the contributions of students in the group projects of 2022, 2023, and 2024, who took part in the assessments and shared their perspectives on the survey. We appreciate the financial support from the Faculty of Science and Engineering at Linköping University, which enabled survey development and implementation.