Analyzing Student’s Failure in VIAT-MAP Reconstruction Process and the Relation to Learning Gain

This study investigates the cognitive dynamics of students' argumentative reasoning in a digital learning environment using the Toulmin Argument Model. It addresses two core issues: (1) whether the interaction between correct and incorrect argument components affects learning improvement, and (2) whether the initial selection order (Ground First vs. Warrant First) moderates this interaction. The importance of this study stems from an inadequate understanding of how students' misconceptions and sound reasoning coexist and influence learning outcomes. Using a pretest-posttest design, examines the impact of the interaction between correct responses (TT_s) and incorrect responses (TF_s) on learning outcomes. Log data capturing each student's decision making process were classified and analyzed using simple linear regression and slope analysis. The findings revealed a significant interaction between TT_s and TF_s, indicating that accurate responses have a stronger effect as students' errors increase, thus supporting the theory of compensatory cognitive processes. Although the order of argument components (FirstDomain) did not influence this effect, the model demonstrated satisfactory statistical validity (R² = 0.386, p = 0.021). These results suggest that the learning strategy should be effective not only in eliminating misconceptions but also in improving sound reasoning, especially for students with weak conceptual understanding. Strengthen the Learning Analytics domain by emphasizing the potential use of log data indicators to assess and support adaptive learning in argument based learning environments.