From Manipulation to Abstraction: The Impact of Flexible Decomposition on Numerical Competence in Primary School
This study examines the effectiveness of a structured instructional approach to decomposition and recomposition of large numbers in six primary school classes (three Year 4 and three Year 5, N = 120) using a quasi - experimental design with a control group. The 12 - week intervention is grounded in the Concrete Pictorial Abstract (CPA) progression. The experimental groups achieved average gains of 34.0 points (Year 4) and 29.6 points (Year 5) out of 100, significantly higher than the control groups (16.4 and 11.1 points; p < .001). The Time Group interaction in the mixed ANOVA reached {\eta}^2p = .931. The ANCOVA with the pre - test as covariate estimated an adjusted difference of 18.25 points (F(1,117) = 2,978.10, p < .001, \eta^2p = .962), confirming the robustness of the effect after controlling for baseline differences. Four-week retention exceeded 97% in the experimental group. Internal reliability of the instrument was satisfactory (Cronbach's {\alpha} = .735).