Beyond the Final Code: Understanding Student Programming Behaviors through Process Journals
Abstract
As students can now use Generative AI to easily solve introductory programming problems, there is a greater need to focus pedagogy on the programming process, not just the final product. Process journals, where students document their problem-solving steps while working on a coding problem, can encourage students to reflect on their learning and make their thinking visible to instructors. We qualitatively analyzed 48 process journals from an introductory programming course to understand student behaviors and interpret these behaviors through the theoretical lens of self-regulation. Our findings revealed ways in which students understand and approach the programming task, how they test and debug their code, how they consulted resources, and how they reflected or managed their learning. We classified these behaviors as positive, negative, or neutral based on how much self-regulation and engagement they showed. Although some students were not able to complete the basic requirements or showed little extra engagement after satisfying the base requirements, most students showed behaviors beyond the minimum requirements with the process journal and some demonstrated thinking beyond solving the problem. In addition, we found that the behaviors students showed were not strongly related to their exam score or their self-reported metacognitive self-regulation metric. Our study presents process journals as a tool that can give educators insight into both desirable and undesirable behaviors of the stages students go through when solving programming problems. This allows educators to stage interventions that support student learning and metacognitive development.