What does process mining of feedback-behavior reveal about problem-solving in chemistry undergraduates?

Feedback is known to play a powerful role in the process of learning and problem-solving. In a problem-solving context where feedback is expected to positively influence the process of achieving the target goal, feedback information can be elicited through feedback-seeking questions. In this study, we analyze different purposes for which feedback is both sought and given by chemistry undergraduate students (n=3) as they solved representational problems concerning a medicinal drug. The students were left unsupervised and, with their consent, were video-recorded to capture their problem-solving process. We used the transcript of the video data and coded students' interactions to identify feedback events concerning 'the task', 'process', and 'regulation' of the problem-solving activity. These events were then used to generate process models of feedback behavior of students, individually and collaboratively, through ProM- a process mining tool. Our models suggest that feedback questions on tasks and processes dominate the beginning of the problem-solving activity, while feedback responses on processes and regulation dominate the end phase. Our work has implications for designing activity problems. We suggest designing well-thought-out focus questions at the beginning of problem-solving exercises to both scaffold and facilitate students' learning because it aligns with the natural unsupervised flow of problem-solving, as is observed in this study.