Development of SpeakUp digital platform based on digital literacy and local wisdom to improve English speaking competence of vocational students
This study developed and preliminarily evaluated the SpeakUp Digital Platform, an interactive English speaking learning platform grounded in digital literacy, Task-Based Language Teaching (TBLT), and Banten local wisdom to improve vocational high school students' speaking competence. A Research and Development (R&D) design using the ADDIE model was implemented at SMK Negeri 10 Tangerang and SMK TI YPML Tangerang, Banten Province, during the 2025/2026 academic year. The platform integrates culturally contextualized speaking tasks, self-recording and playback, vocabulary banks, sentence starters, analytic rubrics, and structured teacher feedback. Needs analysis involving 78 students showed that more than 70% experienced speaking fluency problems, 71% reported vocabulary limitations, 62% faced pronunciation difficulties, and 79% indicated low confidence and speaking anxiety. The product was validated by material and media experts, revised through limited testing with 10 students, implemented in a field trial with 63 students and one teacher, and evaluated through a one-group pretest-posttest design involving 60 students. Expert validation produced a combined feasibility score of 92.00% (very feasible), while practicality assessment yielded an overall score of 88.31% (very practical). Speaking scores increased from a pretest mean of 71.15 (SD = 2.69) to a posttest mean of 91.75 (SD = 1.79), with a significant paired-sample t test result, t(59) = -51.151, p < 0.001, Cohen's d = 6.60, and N-Gain = 0.714. Supplementary PLS-SEM indicated that practice intensity (β = 0.412), learning regulation (β = 0.287), and speaking anxiety management (β = 0.256) were the strongest direct predictors of speaking outcomes (R² = 0.74). Within the constraints of a single-group design, these findings should be interpreted as preliminary evidence of format-level improvement in vocational EFL speaking practice rather than as proof of effectiveness; because no control group was included, the observed gains cannot be attributed causally to the platform, and controlled, longitudinal studies are required to confirm them.