Deep Learning-Based Adaptive Practice Improves Vocabulary Mastery Among Indonesian EFL Learners

Henryani (Pendidikan Bahasa Inggris, Institut Agama Islam Negeri Parepare, Parepare, Sulawesi Selatan)
Zulfah (Pendidikan Bahasa Inggris, Institut Agama Islam Negeri Parepare, Parepare, Sulawesi Selatan)
Abdul Haris Sunubi (Pendidikan Bahasa Inggris, Institut Agama Islam Negeri Parepare, Parepare, Sulawesi Selatan)
Mujahidah (Pendidikan Bahasa Inggris, Institut Agama Islam Negeri Parepare, Parepare, Sulawesi Selatan)
Magdahalena Tjalla (Pendidikan Bahasa Inggris, Institut Agama Islam Negeri Parepare, Parepare, Sulawesi Selatan)

IDEAS: Journal on English Language Teaching and Learning, Linguistics and Literature
Vol. 14 No. 1 (2026): IDEAS: Journal on English Language Teaching and Learning, Linguistics and Lite

This study investigated the effectiveness of Deep Learning-Based Adaptive Practice (DLBAP), an instructional approach grounded in meaningful learning, retrieval practice, spaced repetition, and adaptive reinforcement rather than artificial intelligence-based deep learning, in improving English vocabulary mastery among Indonesian junior secondary EFL learners. The research addressed whether DLBAP was more effective than conventional vocabulary instruction in enhancing students’ vocabulary breadth, vocabulary depth, retention, and productive vocabulary use. The study employed a quantitative quasi-experimental design with a pre-test, post-test, and delayed post-test control group design. The participants consisted of 30 eighth-grade students of MTsS PP Nurul Haq Benteng Lewo, South Sulawesi, Indonesia, divided into an experimental group and a control group. Data were collected through vocabulary tests, classroom observations, and field notes. The experimental group received DLBAP through contextual exposure activities using CoSpaces Edu and adaptive reinforcement using Wordwall, while the control group received conventional instruction. Data were analyzed using descriptive and inferential statistics, including independent and paired samples t-tests. The findings revealed that the experimental group achieved substantially higher vocabulary mastery than the control group, with mean scores increasing from 68.53 in the pre-test to 93.53 in the post-test and remaining relatively high at 89.20 in the delayed post-test. Independent-samples t-test analysis indicated a statistically significant difference between the experimental and control groups in post-test performance (p < .05), with a large practical effect size. The experimental group improved from a mean score of 68.53 in the pre-test to 93.53 in the post-test and maintained a relatively high score of 89.20 in the delayed post-test. The findings indicate that integrating contextual learning, retrieval practice, spaced repetition, and adaptive reinforcement can significantly strengthen sustainable vocabulary learning and support technology-enhanced EFL instruction.