Abstract
Science literacy remains a challenge in secondary education, particularly in scientific domains requiring conceptual understanding, scientific reasoning and real-world application. This study assessed science literacy in electricity, magnetism and alternative energy among urban junior high school students in Indonesia and evaluated the measurement quality of the assessment instrument using the Rasch model. A quantitative descriptive design involved 550 Grade IX students from selected urban schools. Rasch person-fit analysis using infit and outfit mean square criteria (0.50–1.50) identified 61 misfitting respondents, resulting in a final sample of 489 students. The instrument was developed according to the PISA 2018 science literacy framework and analyzed using the Dichotomous Rasch Model. The analysis examined person ability, item difficulty, reliability, item fit, unidimensionality, local independence and measurement precision. Results demonstrated acceptable psychometric properties, with Cronbach’s alpha of 0.81, person reliability of 0.78 and item reliability of 0.85. Unidimensionality was supported by 42.5% variance explained by measures and a first residual contrast eigenvalue of 1.9, while local independence criteria were satisfied. The mean person measure of −0.45 logits indicated that student ability was below the average item difficulty level, suggesting generally low science literacy in the assessed domains. The findings indicate a gap between expected competencies and student performance in fundamental physics and energy-related concepts. The study provides psychometrically grounded evidence for improving science assessment and instruction.
Keywords: Alternative Energy, Electricity, Magnetism, Rasch Model, Science Literacy.