Analysis of Feedback Quality on Engineering Problem-solving Tasks

In this research paper, we examine the types of feedback provided to students on engineering problem solving tasks. The assessment for learning conceptual framework is adopted, which suggests that assessment tasks should further learning rather than being only summative. In this work the feedback observed on marked midterm tests and final exam papers is coded to investigate whether the feedback aligns with an assessment for learning approach. The types of feedback on the papers is characterized using a hierarchical schema with check marks (basic validating feedback) being the least effective, and textual comments (elaborating feedback) being the most effective. The proposed classification is then used to code graded student test papers (naturalistic material) from three electrical engineering courses. The material includes 7 problems from each course, leading to 21 engineering problems in total. Between 16 and 27 graded student solutions are randomly selected for analysis for each problem. The results demonstrate that poor quality student solutions receive less, and less valuable feedback than high quality student work. The results also exhibit a high degree of variability between types of feedback provided on student work. The findings of this study are useful in informing instructional design and changes to assessment practices.