This thesis investigates whether presenting programming assignments in a scaffolded format can improve learners’ performance and experience compared to the traditional formats. A custom-built online experimental platform was developed to deliver scaffolded and traditional versions of automated assessment tools through A/B testing. The platform captured both quantitative metrics (e.g., time on task, number of submissions, pre- and posttest scores) and qualitative feedback (e.g., user perception and preferences). The study involved 40 participants, both beginner and experienced learners, and focused on Python programming assignments related to the loops topic. The results showed that the scaffolded format increased task duration by 70% for beginners, but improved learning experience. Meanwhile, experienced participants showed a greater tendency towards the traditional format because of its simplicity.
Nicholas Jovianto (Wed,) studied this question.