Search papers, labs, and topics across Lattice.
This study investigates how accurately students perceive their partners' use of generative AI tools during collaborative programming assignments and its impact on team performance. By analyzing data from 103 student pairs over three waves, the research reveals that greater misalignment in perceptions of AI use correlates with lower project scores, particularly affecting teams with weaker prior programming skills. Notably, the findings indicate that face-to-face interactions do not significantly reduce perception misalignment, highlighting the need for strategies to enhance transparency in collaborative settings.
Misaligned perceptions of AI use among student partners can significantly hinder collaboration, especially for those with lower programming skills.
Collaborative assignments are a cornerstone of programming education. Effective collaboration during a programming project depends on the formation of reasonably accurate beliefs about how each partner works. Generative AI tools, now widely used by undergraduate students, have introduced a consequential and largely invisible new dimension into collaboration: each student's use of AI. When partners collaborate remotely, they interpret partners' ability and effort through their code. This raises the question of how accurately students perceive each other's AI use in collaborations, and if a misalignment in these perceptions relates to team performance. To address this question, we conducted a three-wave longitudinal study of 103 student pairs in an introductory software engineering course. We found that greater misalignment between partners' beliefs about each other's AI use early in the project was associated with lower final project scores. The effect of such misaligned perceptions is the strongest in teams with lower prior programming performance, suggesting that low performing students pay a higher cost of misaligned perceptions. The perception misalignment does not consistently decrease through face-to-face pair-programming sessions. This suggests that ways to foster transparency may be needed to support student teams in collaborative programming.
