<![CDATA[This study investigates the effect of Claude AI-assisted assessment with a purpose-built adaptive prompt system on reducing mathematics anxiety among university students in Abstract Algebra. Mathematics anxiety—a multidimensional affective construct distinct from general anxiety and from the cognitive barriers addressed in prior research on mathematical representation—constitutes a significant impediment to learning advanced mathematics. The study employed a mixed-methods design with an embedded experimental model involving 26 fourth-semester students of the Mathematics Education Study Program at Universitas PGRI Pontianak during the 2024/2025 academic year. The primary instrument was the Mathematics Anxiety Rating Scale–Revised (MARS-R) adapted into Indonesian and psychometrically validated (Cronbach's α = .91, CFI = .96). The Claude AI system was engineered with six adaptive prompt types—Calming Prompt, Stepwise Scaffolding, Positive Affirmation, Misconception Diagnosis, Metacognitive Reflection, and Progressive Feedback—automatically selected based on detected anxiety profiles. Results showed a highly significant reduction in mathematics anxiety (Z = −4.41, p < .001, r = .86). Mean scores declined from 46.7 (high) to 30.3 (low), a 35.1% reduction. The previously strong negative anxiety–achievement correlation (r = −.67, p < .001) became non-significant post-intervention (r = −.38, p = .054), indicating successful anxiety decoupling. These findings contribute to understanding affective dimensions in advanced mathematics learning and establish an empirical basis for adaptive AI in academic anxiety management.]]>
Copyrights © 2025
