Direct vs. Indirect Listening Assessments: A Comparative Study of Gap-Filling and Cloze Tests on Social Media Using AI

Abstract

This study investigates the effectiveness of direct and indirect listening assessments by comparing two test formats: gap-filling and cloze tests, centered on the topic of social media. The research involved 12 high school students, equally divided by gender, who were administered AI-generated listening comprehension tests. The gap-filling test, used as a direct listening assessment, required students to fill in blanks while listening to a real-time conversation about social media, engaging them in active auditory processing. The cloze test, serving as an indirect listening measure, involved filling in blanks in a written transcript after listening to a similar conversation. Results indicated that the gap-filling test was more effective in promoting cognitive engagement, real-time processing, and inference-making. Additionally, the use of AI to generate and score the tests enhanced the accuracy, fairness, and relevance of the assessments. The findings suggest that gap-filling tests, supported by AI, provide a more comprehensive evaluation of listening comprehension than cloze tests, particularly when assessing real-world language use.