P1-04: Testing The Effectiveness of Student Self-Grading in an Upper Level Statistics Course

By Wilmina Marget & Caroline Bland (John Carroll University)


The purpose of this talk is to describe how self-grading can be implemented and analyse the benefits for both students and instructors. Based on our pilot study with 16 students, self-grading not only has the potential to improve student understanding but also to decrease the burden of grading on instructors. Although the class size was small, because of the use of technology, we believe this approach would extend well to larger class sizes. For testing the effectiveness of self-grading on conceptual understanding, a cross over design was used, requiring students to complete self-grading for some assignments, while receiving traditional instructor feedback on others. Scores on quizzes associated with each assignment were analyzed to measure improvement in student understanding over traditional instructor grading. We found that for some topics, self-grading improved student understanding. For other topics, no differences in understanding were found, but self-grading did not result in a statistically significant decrease in understanding for any topic. Instructor time spent grading was two to three times longer for instructor graded assignments as it was for student self-graded assignments. Students generally preferred self-grading, saying it made them feel more prepared.