A new scoring method looks at how students build a concept map, not only at the map they end up with.
A concept map shows what a student understands by connecting concepts with labelled links. The kit-build framework turns that into a hands-on exercise. Learners receive the concepts and links taken apart from a teacher's map and put them back together. The system then compares the rebuilt map with the original, quickly and without a human scorer.
The catch is that the check only looks at the final map. Two students can arrive at the same map by very different routes. One builds it steadily, cluster by cluster. Another jumps around and fills gaps at random. The product looks identical, but the thinking behind it is not.
The team treated a concept map as a set of sub-maps. Each sub-map is a small cluster of propositions that share one meaning. The assumption is simple. If a learner really understands a cluster, they will build it in one continuous run rather than scattering it across the session.
The evaluation ran with 60 second-year students in an Object-Oriented Programming course. The goal map held 15 propositions across five sub-maps. Students took a pre-test, worked through the activity, took a post-test, and sat a delayed test a week later.
Sub-map scoring correlated with learning gain at 0.711, a strong relationship. Full-map scoring came in at 0.674, which counts as moderate. In a regression model with both measures plus the pre-test score, only sub-map scoring stayed significant, at a p-value of 0.00959.
The practical takeaway is about feedback. A teacher who sees only the finished map reads a correct answer and moves on. A teacher who can see the process spots the learner who assembled the right map without ever holding the whole idea together.
Rismanto et al. (2024). Evaluating the kit-build concept mapping process using sub-map scoring. Research and Practice in Technology Enhanced Learning.
