Abstract
Under the 2017 and 2018 Courses of Study, the criterion-based assessment is implemented in all subjects from elementary school through senior high school to achieve the integration of teaching and assessment based on three criteria: knowledge and skills, abilities to think, make judgment and express oneself, and attitudes toward independent learning. The National Assessment of Academic Ability is designed to assess the first two criteria. To explore whether these two criteria are distinguishable in practice, confirmatory factor analyses (CFA) were conducted for English, Japanese, and Mathematics tests using large-scale response data. The scree tests indicated strong unidimensionality, suggesting the dominance of a single academic ability. However, two-factor CFA models showed statistically better fit and higher proportions of variance explained than the unitary models across all subjects. These findings suggest that while a strong general ability is measured, the intended distinction between the two abilities is also empirically supported.
