Harjono, Woro Sumarni, Sri Susilogati Sumarti, Munzil, Muntholib, Beni Setiawan, Ibrahim Wicaksono, Dimas Gilang Ramadhani
This study explores the strengths and weaknesses of students in complexometric practical exams by employing a multidimensional feedback analysis approach. Complexometric titrations serve as an essential method for assessing students’ practical and representational competencies (macroscopic, submicroscopic, symbolic) in determining metal ion concentrations through titration. By incorporating AI-driven feedback systems, this research aims to provide personalized, real-time insights into students’ performance across multiple dimensions: macroscopic, submicroscopic, and symbolic. Traditional assessments often focus on final outcomes, missing crucial elements such as the students’ understanding of molecular interactions and symbolic reasoning. This research addresses these gaps by combining AI technologies, such as BERT-based feedback systems, with multidimensional analysis to deliver more comprehensive evaluations. Through Canonical Correspondence Analysis (CCA) and t-SNE clustering, the study identifies patterns in student performance, revealing areas where they excel and where they need improvement. The findings suggest that while students may demonstrate proficiency in practical tasks, they often struggle with symbolic reasoning and submicroscopic understanding. This research contributes to the development of more effective educational strategies by leveraging AI to offer targeted feedback and support, ultimately enhancing learning outcomes in complexometric practical exams and similar educational contexts. © The Author(s), under exclusive licence to Springer Nature B.V. 2026.
Faculty of Mathematics and Natural Sciences, Universitas Negeri Semarang, Semarang, Indonesia; Faculty of Mathematics and Natural Sciences, Universitas Negeri Malang, Malang, Indonesia; Faculty of Mathematics and Natural Sciences, Universitas Negeri Surabaya, Surabaya, Indonesia
