Empirical research on cognitive diagnosis of scientific argumentation ability based on the DINA model

This study focuses on the cognitive diagnosis (CD) of scientific argumentation ability. A hierarchical model of six cognitive attributes was developed by expert cognitive analysis, utilizing the Toulmin Argument Model and SOLO Taxonomy theory. The research was conducted in two stages: In the first stage, a 33-item diagnostic test was developed based on Q-matrix theory, and the validity of the cognitive model and Q-matrix was verified using data from a sample of 240 tenth-grade students (M age = 15.6 years, SD = 0.72); and in the second stage, through multi-model fitting comparisons, the DINA model was identified as the optimal model, and students' scientific argumentation ability was diagnosed. The diagnostic results demonstrated that the construction of the cognitive model and Q-matrix was valid and the DINA model could effectively diagnose the cognitive structure of scientific argumentation, further constructing students' learning paths and providing empirical evidence for the targeted design of scientific argumentation thinking training. These outcomes transform general ability assessment into diagnosis of specific cognitive components, offering targeted evidence for instructional intervention. Overall, the study provides a reliable theoretical and empirical foundation for the assessment and improvement of scientific argumentation competency.