Impact Of Grouping Strategy on Students' Performance and Satisfaction with Teamwork

Practice Paper Recommended citation: Garcia Souto, M. P., Nasrollahy Shiraz, A., Odunsi, A., & Siefker, J. (2025). Impact Of Grouping Strategy on Students’ Performance and Satisfaction with Teamwork. 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.17631643. 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. IMPACT OF GROUPING STRATEGY ON STUDENTS’ PERFORMANCE AND SATISFACTION WITH TEAMWORK M.P. Garcia Souto a, 1 , A. Nasrollahy Shiraz b, A. Odunsi c, J. Siefker d a University College London (UCL), London, UK, 0000-0002-8222-8329 b University College London (UCL), London, UK, 0000-0003-1732-1401 c University College London (UCL), London, UK, 0009-0009-8321-9406 d University College London (UCL), London, UK, 0009-0009-9599-0337 Conference Key Areas: Engineering skills, professional skills, and transversal skills; Continuing education and life-long learning in engineering Keywords: Group work, teamwork, grouping strategies, students’ satisfaction, skills ABSTRACT Competency-based grouping strategies in team projects have not been sufficiently studied in the literature. In this pilot study, we investigated the impact of three different grouping strategies in four team-based student projects in UCL Faculty of Engineering Sciences. These three grouping strategies, all instructor-assigned, are based on setting groups such that they are (i) balanced by having similar average prior academic performance in all teams, (ii) individuals in a team have similar prior academic performance, or (iii) individuals in a team are similar with respect to a combination of academic performance and teamwork performance. In this study we evaluated the impact of these grouping strategies on the group output, as quantified by the group mark, and the team dynamics, as quantified by the Individual Peer Assessment of Contribution (IPAC) score. Surprisingly, the balanced grouping strategy did not necessarily lead to similar group outcomes but generally led to better team dynamics. Overall, we observed that the student perception of the team dynamics does not necessarily match the team’s output. On the other hand, there was an understandable correlation between past performance indicators and the group outcome in similarperformance grouping strategies. This paper concludes by providing a summary of the observed pros and cons for the different grouping strategies. These results are of interest to all engineering educators who implement team-based learning within their curriculum. 1 Corresponding Author M.P. Garcia Souto [email protected] 1 INTRODUCTION The formation of student groups is critical to the success of team-based learning in higher education. Teams must be formed with a proactive, supportive approach, considering factors such as the diversity of skills and backgrounds, to enable effective collaborative working, and avoid common issues like ‘grouphate’ and social loafing (Francis et al, 2025). At UCL, the Individual Peer Assessment of Contribution (IPAC) to teamwork methodology allows the learners to assess the level of contributions of each of their peers (academic and team based), including themselves, from which an IPAC factor is generated after moderation by the tutor. The IPAC factor can then be used in combination with the team mark to generate individual marks for each team member, that reflects the achievements of the team as well as the individual contributions (Garcia-Souto et al, 2019; Garcia-Souto et al, 2019; Garcia-Souto et al, 2024; Garcia-Souto et al, 2024). UCL has developed an LTI that allows practitioners to implement this assessment methodology easily and time efficiently within any institutional virtual learning environment, and it is available to other institutions (email corresponding author if interested on using it). The main three approaches to the formation of student groups namely, self-selected groups, instructor-assigned teams and hybrid strategies that combine elements of both self-selection and instructor assignment, are put forward in literature – each with their merits and demerits (Chang & Brickman, 2018; Race, 2019; Francis et al, 2025). In a self-selection approach, students are given the autonomy to form their own groups, which can foster a sense of camaraderie and ease of communication. However, this method may lead to homogeneous groups in terms of skills or experience, potentially excluding less connected or new students. In contrast, an instructor-assigned approach involves the educator deliberately composing groups to ensure a balance with regards to some characteristic (skills, personalities, personal characteristics or academic abilities), thereby creating heterogeneous teams that are better equipped to tackle the challenges inherent in group work, such as social loafing and uneven participation. Hybrid strategies involve initial instructor-led formation to establish diverse, balanced teams, followed by opportunities for students to adjust or refine their group composition. This work explores three variations of the instructor-assigned team strategies, namely, (A.) balanced by prior academic performance, (B.) similar prior academic performance and (C.) similar prior academic and teamwork performance. The key objectives of this work are: 1.) To examine if students’ previous academic performance is a good basis for predicting how students perform in group activities, 2.) To examine what influences, if at all, the student group formation strategy has on performance in group activities, 3.) To examine if considering previous teamwork quality (i.e., as evaluated by IPAC scores) in combination with students’ academic standing, when forming new groups for a different team activity, has any effect on the outputs of the team activity. 2 CONTEXT AND PRACTICAL WORK 2.1 Grouping strategies and case studies A total of 4 engineering team projects were selected for their grouping strategies, as indicated in Table 1 along with a short description of each. All team projects are part of UCL Integrated Engineering Programme (IEP) degrees, although running within their own disciplines. Three grouping strategies were studied in this paper, namely: • Grouping A – Balanced by prior academic performance: The average mark across all modules from the previous year was used as the past performance indicator, and teams were set up such that each team had a similar past performance indicator. • Grouping B – Similar prior academic performance: The average mark across all modules from the previous year was used as the past performance indicator, and teams were set up such that members of a given team had similar past performance indicators. • Grouping C – Similar prior academic and teamwork performance: The average mark across all modules from the previous year, current year team project marks, and current year IPAC score in earlier teamwork project were used in tandem to form an index we refer to as the J-index here for each student. Teams were then set up such that students within a team had a similar J-index. In case II, teams were initially formed using grouping approach A; however, groups in which all team members were then identified to have a similar past academic performance were re-classified under grouping approach B in this study. Table 1. Case studies summary Case Engineering discipline Year study Academic year Class size Team size Project length Weight (%) × ECTS Grouping strategy I Electronic & Electrical Eng 2 2024/25 118 4-5 1week FTE 15%×7.5 A II Biomedical Eng 3 2019/20 35 in total 4-5 6 months 80%×15 A (17 students) B (18 students) 4 2021/22 90% ×22.5 III Chem Eng 3 2024/25 116 5-6 3 months 45% × 7.5 B IV Chem Eng 2 2024/25 159 5-7 1week FTE 15% × 15 C 2.2 Analysis The analysis included various factors of interest as listed below: - Prior performance of the student, i.e. the prior year academic average or the J-index depending on which one was used - Group mark for current team activity - Individual IPAC scores for current team activity - Type of grouping as defined in section 2.1. - Average and standard deviation for prior performance within the team - Average and standard deviation for IPAC scores within the team - Individual project mark, calculated as group mark x IPAC score; these are just prospective marks given as indication since they do not consider IPAC tutor moderations. - Delta mark for each student, representing the difference between the current individual project mark and their prior performance score - Delta mark by IPAC, which is the difference in mark a student gets for the project when their individual IPAC is considered (i.e. individual mark – group mark) We investigated the relationship between prior and current performance both at group and individual level for the different grouping strategies (Fig. 1 and 2). We investigated the cohesiveness within the teams by means of the standard deviation of IPAC scores within the team for each grouping strategy (Fig. 3) and checked if it was related to the overall group mark achieved by the team (Fig 6.). Furthermore, we looked at how the individual marks change with respect to their prior performance (Fig. 4) and with respect to the group mark by means of their individual IPAC score (Fig. 5), and the effect of the grouping strategy. When relevant, statistical analysis was performed to understand if distributions were statistically different at 0.05 level. We consider data to be parametric, i.e. data is measured on a scale with the differences between the values being meaningful and consistent, data follows normal distribution, and meet homogeneity of variance. Therefore, we used One-Way ANOVA test. Homogeneity of variances were corroborated with Tukey tests. 2.3 Results and Discussion Considering Fig. 1, an interesting observation associated with grouping approach A is that having balanced groups based on the past academic performance alone does not necessarily lead to similar group performances as a relatively wide range of group marks is observed. On the other hand, in approaches B and C, there is a direct correlation between the average prior performance of individuals in a team and how the team performs, as measured by the group mark. In this case, the slope of the best fit line for the data in Fig. 1 is closer to one for approach C in which the prior performance metric is extended beyond considering only the academic performance (approach B), to also include the previous teamwork quality. The individual student marks (i.e., the group mark adjusted by the IPAC score), as shown in Fig. 2, show little to no correlation with the associated past performance indicator for grouping approach A. Similar to the group mark, the slope of the best fit line is greater in Fig. 2 for approach C compared to that for approach B. This may indicate that the combination of past academic and teamwork performances may serve as a better indicator of the current individual performance compared to the case in which only the past academic performance is considered. However, it is important to note that the average past academic performance of the team may play a role in predicting the individual mark when members have a similar past academic performance indicator. The lack of this correlation for balanced teams should be highlighted here again as one would have expected the high performing students to achieve a high IPAC score assuming that they would continue to work harder in the team on average. However, this is not always the case, and it might be e.g. that they adjust their effort to the lower team’s achievement expectations or are not equally skillful at teamwork. Beyond group performance, as shown in Fig. 3, students appear to be more satisfied with the team dynamics in a balanced grouping approach although the team may not necessarily perform well. This is inferred since the mean of the standard deviations of the IPAC scores is lower in this grouping approach compared to those in grouping approaches B and C. We can hypothesise that in grouping approach A a natural hierarchy forms where consensus is easily reached over the individual roles. However, in other approaches, team members may clash more given that a natural hierarchy in not necessarily established if all team members have a similar past academic performance; thus, consensus over individual roles may not be as easily reached. Therefore, grouping approaches B and C may be more prone to the deterioration of the team dynamics. The effect of this can also be seen on the data in Fig. 4 where in most cases for grouping approach A, the IPAC score has only a small impact on the deviation of the individual mark from the group mark. In contrast, as also shown in Fig. 4, the IPAC score in grouping approaches B and C can lead to a more significant deviation of the individual marks form the group mark. As shown in Fig. 5, grouping approach 1 may provide an opportunity for those with normally lower academic performance indicators to excel in the group activity. This is seen in the net positive change in the average individual mark compared to the past academic performance. Interestingly, grouping approach B shows a net negative change and, understandably, grouping approach C shows no change on average as the past performance indicator already includes the teamwork quality. It should be noted that the average values are very small (close to zero), but the standard deviation is large. This strongly shows that students’ academic achievement and teamwork ability should be consider independently. And if that is the case, are we assessing all the “relevant” elements in group projects or should the assessment have a larger weighting towards teamwork skills? Looking at the grouping strategy, only the distributions for groupings A and B are significantly different at 0.05 significance level. There has been a recurring theme in our results that the perception of students regarding their teams’ dynamics does not necessarily correlate with what the team achieves. In other words, students may be content with how the team members work together even if they do not achieve the defined goals. This can be further observed in Fig. 6, noting that a similarly wide range of group marks have been achieved despite different IPAC score standard deviations (ANOVA shows no significant difference at 0.05 level) even though for the largest IPAC standard deviation category (>0.15) we note an average drop of ~5%. Best fit lines: Grouping B: y = 22.49 + 0.64x; R2 = 0.720 Grouping C: y = 15.55 + 0.77x; R2 = 0.640 Fig.1. Performance of the team, prior vs current. Red line represents the identity line (y=x) Best fit lines: Grouping A: y = 64.42 + 0.08x, R2 = 0.005 Grouping B y = 23.69 + 0.62x; R2 = 0.468 Grouping C: y = 16.08 + 0.76x; R2 = 0.524 Fig. 2. Performance of an individual, prior vs current. Red line represents the identity line (y=x) Fig. 3. IPAC standard deviation within a team by grouping type. No statistical difference found at 0.05 significance level. Fig. 4. Impact of the IPAC on student’s marks, i.e. individual mark – group mark , or (IPAC – 1)*group mark Fig 5. Difference between current individual project mark and prior performance. Statistical difference found at 0.05 significance level between grouping A and B. Fig. 6. Group performance vs “group dynamics” (or IPAC variability within the group). ANOVA shows no significant difference at 0.05 level. 2 CONCLUSIONS AND IMPLICATIONS Three different grouping strategies have been investigated. This is a pilot study and as such it is limited in scope, but it is primarily aiming to set up the basis for a larger study. However, it has already shown interesting effects of different grouping strategies and questions that are of interest to the general engineering education community. These are: • The pros and cons for each of the grouping strategies are summarized in Table 2. • Using grouping strategy A (balance of teams by prior academic performance) does not necessarily lead to similar outcomes by the teams which was unexpected. • Traditional academic performance and teamwork performance are not necessarily related, and if so, why are we placing such a big emphasis on the group’s output, should we be training and assessing more strongly students in their capacity to work in teams? • The student perception of the team work dynamics does not correlate with the quality of the team’s output. Some of the limitations we note of the current study are that: (i) limited data; (ii) effect of co-factors such as dependency with module assessment details, year of study with inherent increase in expectations in later years, etc.; (iii) effect of other individual personal factors at play which are not considered here, so we can only speak in statistical terms. We plan to address these in future work but extending the study. The IPAC (Individual Peer Assessment of Contribution to group work) assessment methodology has been identified as a way to assess the ability of students to do teamwork and their level of contribution to the team, and even identify groups with poor group dynamics. Staff from all universities are invited to try the IPAC system, that makes the running of this assessment methodology very easy - just send email to Pilar Garcia Souto ([email protected]; [email protected] ). Table 2. Summary of grouping strategy pros and cons Grouping Pros Cons A • Potentially better team dynamics, possibly by natural role hierarchy • Facilitates peer-to-peer academic support and growth opportunities for students with weaker past academic performance • Less likely for the students with weaker past academic indicators to lead activities, and tendency for the weaker students to look up to their stronger peers. B • Incentivise students with weaker past academic indicators to take responsibility and ownership for team project • Giving the students with weaker past academic indicators the opportunity to lead and manage the project • Opportunities for groups with high past academic indicators to work at the highest standards and challenge themselves. • Less robust in-group academic support mechanisms for the students with weaker past academic indicators • Potentially more prone to the deterioration of the team dynamics across the board C • Giving the students with weaker past academic and teamwork indicators (J-index) the opportunity to lead and manage the project • Places emphasis on the difference between academic performance and teamwork ability • Incentivise students with weaker Jindex to proactively engage with their team • Less robust in-group academic and teamwork support mechanisms for the students with weaker past performance indicators • Potential break down of teams with low J-index 3 ACKNOWLEDGEMENTS Thanks to the UCL Centre for Engineering Education for creating a culture of research on engineering education that has brought the authors together, and for their financial support to attend the SEFI conference.