Innovative AI-Driven Approaches to Mitigate Math Anxiety and Enhance Resilience Among Students with Persistently Low Performance in Mathematics
Georgios Polydoros, Victoria Galitskaya, Pantelis Pergantis, Athanasios Drigas, Alexandros Antoniou, Eleftheria Beazidou
Abstract
This study explored innovative methods for teaching mathematics to seventh-grade students with persistently low performance by using an AI-driven neural network approach, specifically focusing on solving first-degree inequalities. Guided by the Response to Intervention (RTI) framework, the intervention aimed to reduce math anxiety and build academic resilience through the development of cognitive and metacognitive strategies. A rigorous pre- and post-test design was employed to evaluate changes in performance, anxiety levels, and resilience. Fifty-six students participated in the 12-week program, receiving personalized instruction tailored to their individual needs. The AI tool provided real-time feedback and adaptive problem-solving tasks, ensuring students worked at an appropriate level of challenge. Results indicated a marked decrease in math anxiety alongside significant gains in cognitive skills such as problem-solving and numerical reasoning. Students also demonstrated enhanced metacognitive abilities, including self-monitoring and goal setting. These improvements translated into higher academic performance, particularly in the area of inequalities, and greater resilience, highlighting the effectiveness of AI-based strategies in supporting learners who struggle persistently in mathematics. Overall, the findings underscore how AI-driven teaching approaches can address both the cognitive and emotional dimensions of mathematics learning. By offering targeted, adaptive support, educators can foster a learning environment that reduces stress, promotes engagement, and facilitates long-term academic success for students with persistently low performance in mathematics.