Advancing higher education students’ assessment experiences with conversational agents
摘要
Conversational agents have been designed to improve instruction quality and support student learning. In addition to their instructional use, they can be incorporated into assessment—conversation-based assessment (CBA). This study primarily introduces a CBA with selected-response and constructed-response tests as a formative assessment tool for higher education students, while also reporting student attitudes. CBA was designed using Rasa and deployed to Google Chat for students in two sections of an undergraduate-level course to use. Results show that CBA with both formats produces high standard performance measures and confidence scores for each student response. Most students reported positive attitudes toward CBA via the survey. Overall, accurate dialogue moves within CBA underscore its effectiveness in measuring higher education students’ knowledge and skill, while positive student attitudes toward CBA indicate the promise of conversational agents to enhance student assessment experiences by providing a more interactive assessment environment.