Supporting the Semi-automatic Feedback Provisioning on Programming Assignments

Feedback is important for learning, however, manual feedback provisioning is time and resource consuming. For programming education, various systems have been developed to automate recurring tasks or the whole feedback generation. As fully automated systems are not always a viable option, this paper investigates how teaching assistants can be supported in semi-automated assessment of programming assignments to give good feedback more easily. An existing semi-automated e-assessment system is extended with configurable feedback snippets as well as adaptively recommended feedback snippets based on feedback of similar submission in order to evaluate the effects on the grading and the feedback given by teaching assistants. The results indicate that such feedback snippets lead to more consistent and motivational feedback, can help finding mistakes, and have no impact on the awarded grades.