How computational thinking can be integrated in statistical learning: A cuboid framework

Abstract

In the context of an increasingly data-intensive society, the integration of Computational Thinking (CT) into statistics education is essential to prepare students with the analytical and problem-solving competencies required for navigating complex data environments. Despite growing recognition of its importance, existing pedagogical practices frequently lack systematic didactical frameworks that effectively embed CT within statistical learning, particularly in higher education. Addressing this gap, the present study introduces a novel hypothetical didactical design—termed the Cuboid Framework—which systematically integrates CT components into the learning of descriptive statistics using the R programming language in a Google Colab environment. This research employed the Didactical Design Research (DDR) methodology, emphasizing the prospective and metapedadidactic stages to construct and evaluate the framework. Targeted at third-semester undergraduate students enrolled in an introductory statistics course, the Cuboid Framework aligns with learners’ developmental levels in both statistical reasoning and CT proficiency. The model is organized as a 5 × 4 × 4 structure, comprising five core statistical tasks, four structured didactical situations (action, formulation, validation, and institutionalization), and four CT elements (decomposition, pattern recognition, abstraction, and algorithmic thinking). Validation procedures included expert review through focus group discussions (FGDs) and an initial classroom implementation followed by metapedadidactic analysis. Findings reveal that the Cuboid Framework fosters a coherent learning progression, enhances students’ engagement in statistical inquiry, and supports the development of CT competencies. Classroom observations confirmed that the intentional design of didactical situations facilitates students’ cognitive adaptation to computational tasks. While preliminary analyses indicate strong theoretical and practical coherence, further retrospective studies and quantitative evaluations are necessary to ascertain the long-term effects on student learning outcomes. This study contributes a structured and theoretically grounded model for CT integration in statistics education, with implications for improving curriculum design and instructional practice in mathematics education. Future research should aim to test the scalability and efficacy of the Cuboid Framework across diverse educational settings.