SQL Island for SQL Learning in Indonesian Vocational Education: A Quasi-Experimental Study of Learning Outcomes and Motivation

Abstract

Game-based learning (GBL) has gained considerable attention as a strategy to enhance student engagement in technical subjects, yet evidence from vocational secondary school contexts in Southeast Asia remains limited. This study examined the effect of SQL Island — a web-based SQL learning game — on the learning outcomes and motivation of Grade XI vocational students (RPL program) at SMK Negeri 2 Karanganyar, Indonesia. A quasi-experimental Pretest-Posttest Control Group Design was employed with 53 students (n=23 experimental, n=30 control) selected through cluster random sampling. Instruments included a 12-item essay test (ICC=0.985) and a 22-item Likert scale motivation questionnaire (Cronbach's α=0.770), both validated using Aiken's content validity formula. Data were analyzed using ANCOVA, Independent Samples t-test, Mann-Whitney U, and N-Gain in Jamovi. Results showed no significant effect of SQL Island on learning outcomes (ANCOVA: F=0.465, p=0.498, η²p=0.009) or motivation (t=0.030, p=0.976). Post-hoc power analysis revealed the study was substantially underpowered (1−β=0.07), contextualizing the null result. Critically, a ceiling effect in the experimental class (mean pretest=87.2/100) severely constrained measurable improvement, yielding negative mean N-Gain scores for both groups (experimental: −0.441; control: −0.186). These findings contribute methodological insight on ceiling effect as an underreported confound in quasi-experimental GBL research, and offer practical guidance for future implementations: pre-assessment screening, extended treatment duration (≥4 sessions), and language scaffolding are recommended before deploying SQL Island in vocational settings.