Question Generation Workflow: Incorporating Student-generated content and Peer Evaluation

Abstract

Most classroom settings offer limited opportunities for students to engage with the class material. Furthermore, students are not able to receive timely feedback on their work due to limited number of staff who are able to provide comments. With increased student work to grade, the time it takes for staff to grade increases. In this thesis, we propose a question-generation workflow that addresses these issues. In this workflow, students have the opportunity to learn class material more in depth by creating their own multiple-choice questions (MCQs) on class material. As part of the workflow, students also answer and leave feedback on MCQs written by their peers. Both the student-written question and the feedback on those questions go to staff for review and staff can rely on the feedback to determine more quickly the quality of a student-written question. The application of this workflow with Questionable, a web application that can handle both student and staff needs as part of this process, in a computer science course at MIT has shown promising results of students engaging with the course material and improved grading experience for staff.