Where students struggle: Using data to uncover bottleneck in chemistry literacy

Abstract

Understanding students' struggles in learning chemistry requires a multidimensional perspective that considers not only content mastery but also application, reasoning, and affective factors. This study aimed to identify bottlenecks in students' chemical literacy by analyzing performance across four key dimensions: content knowledge, contextual application, higher-order thinking skills (HOTS), and affective aspects (motivation and attitudes). Using a data-driven quantitative approach, assessment scores on electrochemistry topics were examined through advanced educational visualisations. Findings reveal that while students demonstrated strong performance in HOTS and moderate understanding of content, their ability to apply concepts in real-life contexts was limited. The most significant bottleneck emerged in the affective domain, where low scores were consistent and widespread. Correlation analysis further indicated weak relationships between dimensions, particularly the contextual aspect, which appeared conceptually independent. Visual tools such as Sankey diagrams highlighted progressive performance decline from content to affective domains. These results suggest that affective factors play a critical role in learning outcomes and require targeted intervention. The study underscores the importance of integrating data science in educational analysis to capture nuanced learning patterns and guide personalized instructional strategies.