Concept maps have been used to assess knowledge acquisition, track learning, and reveal mental models. This study proposes and validates a computationally measurable coding scheme to overcome educational assessment limitations: time consuming, inconsistency in coding, and difficulty in measuring semantic and structural complexity. The three-step coding scheme includes (i) classifying vertices and edges using a guidebook, (ii) training coders through a standardized manual, and (iii) validating reproducibility via inter-rater reliability (IRR) using Fleiss' Kappa. Results from 22 undergraduate researchers coding six student-generated maps yielded moderate to substantial agreement (Kappa = 0.67 for vertices, 0.45 for edges), supporting both the accuracy and consistency of the scheme. This coding scheme enables scalable, real-time analysis of student thinking and lays the groundwork for automated feedback systems, with potential applications in adaptive learning and tracking of engineering identity in students pursuing engineering education. Beyond enabling structural and semantic analysis of student thinking, the scheme supports automated translation of hand-drawn maps into a digital, analyzable format laying the groundwork for scalable, real-time feedback systems in engineering education. By aligning methodological precision with reflective assessment practices, this research introduces a tool for measuring how students organize and evolve their understanding within project-based curricula. Future directions include integrating AI for automated coding, with promising applications in adaptive learning environments, formative feedback mechanisms, and long-term identity tracking in engineering programs.
原文链接
分享
引用
