國家教育研究院

In this study, the effects of Taiwan's 2019 competency-based curriculum reform on students' mathematical literacy are examined using a difference-in-differences (DiD) approach with data from PISA 2018 and 2022. The reform's staggered rollout generated differential exposure across cohorts: Grade 9 students experienced approximately two to three years of implementation at the lower secondary level (ISCED 2), whereas Grade 10 students encountered less than one year of exposure at the upper secondary level (ISCED 3), providing a quasi-experimental setting for comparative analysis. Descriptive analyses based on Multidimensional Item Response Theory (MIRT) show that Grade 9 students achieved consistent improvements across PISA mathematics subscales (d ? 0.24–0.27), particularly in problem solving and reasoning, which align with the inquiry-oriented and competency-focused objectives of the reform. In contrast, Grade 10 students exhibited smaller performance improvements (d ? 0.08–0.12). The DiD estimates indicate that the Grade 9–Grade 10 performance gap widened by approximately 15 points (?0.15 SD) between 2018 and 2022, reflecting larger relative gains among cohorts with longer exposure to the reformed curriculum. These results remain robust after controlling for socioeconomic status (ESCS) and pandemic-related school disruptions. Equity analyses suggest a modest tendency toward a steepening of the ESCS gradient, with early gains appearing more concentrated among socioeconomically advantaged students. Overall, the findings provide early quasi-experimental evidence that prolonged exposure to a competency-based mathematics curriculum is associated with higher levels of mathematical literacy, while underscoring the need for sustained, equity-oriented supports to promote inclusive and long-term learning benefits.