The Future of Assessment: Automating Multiple Choice Question Generation

Creating multiple-choice questions (MCQs) automatically from the text has gained prominence as a research field due to the broad acceptance of MCQs for large-scale assessments in many areas. Despite their utility, manual MCQ generation is both costly and time-consuming. Consequently, since the late 1990s, researchers have been increasingly drawn to automatic MCQ generation. Numerous systems have been developed in this pursuit, prompting a systematic review in this paper. Our study presents the outcomes of this review, along with a structured workflow encompassing phases for automatic MCQ generation. In each phase, we explore and discuss the methods found within scholarly works. Additionally, assessment techniques employed to gauge the standard of MCQs produced by the system are examined. Finally, topics for future research aimed at enhancing the existing literature are identified.