Impact of Team Formation Type on Students’ Performance in PBL-Based Software Engineering Education

Abstract

In Requirements Engineering (RE) Courses, it is a common teaching practice to adopt a Problem-Based Learning (PBL) approach in which the students are divided into teams to solve problems. These teams can be defined according to different criteria and evaluated using performance assessment models. This paper investigates the impact of team formation type on students’ performance in PBL-based software engineering education in RE courses. The study analyzes 25 teams across five postgraduate RE courses conducted in 2022 and 2023 using a mixed approach (qualitative and quantitative). In three of these courses, the students self-selected the teams (S); in the other two, a team formation method (TFM) was used. We analyzed how performance assessment results and project scores differ between self-selected and TFM-formed teams. We also explored how performance in some soft skills varies between assessments or team formation methods. The difference in average performance between S and TFM teams is statistically significant. We cannot conclude that there is a statistically significant difference in grades between the S and TFM teams. Interestingly, we also observed that the impact of the type of team formation is relatively stable, regardless of the assessment over time.