The current paper represents a study on how team size and composition (level of prior acquaintance between team members) affect students' performance and psychological comfort working on computer science projects in the online environment. In the conducted study, the students collaborated using online methods, as the coronavirus pandemic limited human interaction. After working on two separate projects during the Software Project Management course, while being members of both self-elected and randomly assigned teams, the students have been asked to complete an anonymous form that assesses the performance of their team by four metrics - coordination, quantitative indices, quality, and psychological comfort. 44 students have completed the form, and the results suggest that randomly assigned teams have significantly decreased psychological comfort compared to the self-selected ones. The students also felt a similar reduction in the number of tasks performed and the quality of their solution, although their final grades suggested a less significant drop in performance. Also, the increase of team members has determined an increase in the difficulty of coordination between members and the tendency of unequal distribution of tasks. As future work, we think that psychological profiling might give an in-depth research and better understanding for the process building a team's size and componence.

team performance, online teaching, level of acquaintance, team size, collaboration in software development