Developing and Evaluating an Arabic Speaking Curriculum in an Indonesian Islamic Boarding School: A 4-D Model-Based Quasi-Experimental Study
Abstract
Arabic curricula for non-native learners may include all four language skills while providing insufficient structure for sustained oral production in classroom implementation. This study developed and evaluated speaking-focused instructional units for first-year middle-school students at Al-Wadi Al-Mubarak Institute, Bogor, Indonesia. Research and development followed the define, design, and develop stages of the Thiagarajan 4-D model and was combined with a nonrandomized quasi-experimental pretest-posttest control-group design. The available field sample comprised 32 students, with 16 in each group; no documented a priori power analysis was available. Data were collected through interviews, observation, curriculum analysis, expert-validation forms, a speaking test, and an analytic oral-performance rubric. The surviving study records do not document independent multiple rating or an inter-rater reliability coefficient; this measurement limitation is therefore acknowledged. The product comprised speaking units for Part One of Arabic Between Your Hands in a student book and teacher guide, integrating communicative language teaching, task-based learning, cooperative work, role-play, formative assessment, self-assessment, and portfolios. The experimental-group mean increased from 9.75 to 17.19, compared with 9.38 to 10.63 in the control group. The adjusted posttest difference was 6.272 points (95% CI [5.688, 6.856]); ANCOVA yielded F(1, 29) = 481.853, p < .001, partial η² = .943. This effect size describes variance associated with group membership after adjustment within this sample, not the percentage of improvement caused by the curriculum. The study offers a methodological and contextual contribution by operationalizing needs analysis, the 4-D process, communicative tasks, and performance assessment within a boarding-school curriculum. Findings remain limited by the small, single-institution, nonrandomized sample and undocumented inter-rater reliability.