Enhancing Metacognitive Listening and Learner Autonomy Through an Artificial Intelligence-Mediated Metacognitive Listening Textbook: A Quasi-Experimental Mixed-Methods Study

Listening instruction in language education often remains product-oriented, focusing on comprehension accuracy rather than awareness of cognitive processes and learner self-regulation. Grounded in metacognitive and self-regulated learning frameworks, this study examined the effectiveness of an artificial intelligence (AI)-mediated metacognitive listening textbook designed to promote metacognitive awareness and learner autonomy through guided reflection and adaptive feedback. The textbook implemented a four-phase metacognitive cycle (planning, monitoring, evaluating, reflecting) supported by AI feedback that helped learners diagnose comprehension challenges, plan strategies, and monitor progress. A 12-week quasi-experimental mixed-methods design was conducted with 500 upper-secondary learners divided into experimental ( n = 250) and control ( n = 250) groups. The intervention comprised 24 classroom sessions featuring culturally situated informational-listening tasks aligned with learners’ communicative experiences. Pre- and post-intervention measures included a standardized listening comprehension test, the Metacognitive Awareness Listening Questionnaire (MALQ), the Self-Regulated Learning Questionnaire (SRLQ), and the Learner Autonomy Scale (LAS), complemented by reflective journals and semi-structured interviews. Analysis of covariance showed that the experimental group outperformed the control in listening ( p < .001, d = 0.91), with higher descriptive gain (+12.4 vs. +4.5) and a greater share scoring ⩾80% (52% → 84% vs. 51% → 65%). Questionnaire results favored the experimental group (MALQ Δadj = 0.38, SRLQ Δadj = 0.41, LAS Δadj = 0.22; all p < .001), with the strongest subscale gains in planning/evaluation, problem-solving, forethought, and strategic time management. Qualitative analyses of 2,250 journals and 36 interviews revealed purposeful planning, real-time monitoring, and reflective repair supported by culturally familiar contexts that enhanced authenticity and motivation. The study proposes a pedagogical model integrating metacognitive instruction, AI-supported reflection, and cultural contextualization as a practical pathway for embedding AI within textbook design, guiding listening pedagogy toward a more strategic, self-regulated, and meaningful learning process.