Automatic distractor generation for multiple-choice English vocabulary questions

Area of Science:

• Natural Language Processing
• Educational Technology
• Second Language Acquisition

Background:

• Automated systems can alleviate teacher workload in language education.
• Developing effective distractors for multiple-choice questions is challenging.
• Current methods for distractor generation may lack quality and efficiency.

Purpose of the Study:

• To propose and evaluate a novel method for automatically generating distractors for English vocabulary multiple-choice questions.
• To enhance the efficiency and quality of educational assessment tools.
• To reduce the burden on human teachers and test creators.

Main Methods:

• Utilizing new sources for distractor candidate collection.
• Employing semantic similarity and collocation information for ranking candidates.
• Evaluating generated distractors through learner testing and expert judgment.

Main Results:

• The proposed method generated fewer problematic distractors compared to a baseline approach.
• AI-generated distractors demonstrated quality comparable to human-created distractors.
• Expert evaluation confirmed the effectiveness of the automated distractor generation.

Conclusions:

• Automated distractor generation is a viable and effective approach for English vocabulary assessments.
• The proposed method offers a promising solution for reducing teacher workload and improving assessment quality.
• Further research can explore expanding this method to other language learning contexts.