Mining and modelling the cognitive strategies students use to construct higher quality causal maps

Abstract

This study examines the cognitive strategies students employ when constructing causal maps and investigates whether specific processing methods - backward versus forward and breadth-first versus depth-first - affect map quality. Building on prior research that linked backward and depth-first processing to stronger argument maps, the study applied similar analytical tools to causal mapping, a task with distinct cognitive demands. Using jMAP software, 37 causal maps were created by students at a Southeastern university in the United States. Within the first five nodes placed, students used backward, forward, breadth-first, and depth-first processing an average of 0.81, 1.84, 1.73, and 1.95 times, respectively, producing maps with a mean score of 125.82 (SD = 55.13) out of a possible 340. Regression analysis revealed that neither backward nor breadth-first processing significantly predicted map scores, though a non-significant trend suggested that more frequent backward processing may support higher-quality maps, consistent with prior findings on argument mapping. While not definitive, these results carry instructional implications: encouraging backward processing and designing diagramming tools that facilitate this strategy without inadvertently biasing students toward less effective approaches. Ultimately, the study highlights that predominant reliance on forward, backward, or breadth-first processing does not necessarily yield higher-quality causal maps as it does in argument mapping, underscoring the importance of understanding task-specific demands to inform instructional design and the development of tools that better support student reasoning and problem-solving.

https://doi.org/10.34105/j.kmel.2026.18.019