<![CDATA[Scientific reasoning in chemistry involves the ability to apply conceptual knowledge in problem-solving, as well as to evaluate issues within broader social, ethical, and environmental contexts. However, conventional assessments often fail to capture this multidimensionality by reducing performance to a single final score. This study uses an integrated learning analytics approach to analyze students’ reasoning performance across two core domains of chemistry learning—applied reasoning and socio-chemical reasoning. A quantitative descriptive design was employed, involving 56 pre-service chemistry teachers who completed four open-ended essay questions, two in each reasoning domain. Student responses were scored using an analytical rubric assessing conceptual accuracy, logical coherence, and justification relevance. Data were analyzed using single-domain and multicomponent strategies, including quadrant profiling, trajectory mapping, clustering, and distribution analysis. Visual tools such as radar charts, spaghetti plots, contour density plots, and alluvial diagrams were used to depict students’ reasoning profiles. Results revealed that most students demonstrated moderate reasoning abilities, although notable inconsistencies were observed between the domains. Individual trajectories exhibited non-linear variations, highlighting diverse cognitive patterns. Clustering and heatmaps indicated distinct learner segments, while alluvial diagrams illustrated transitions between reasoning levels across domains. These findings suggest that students’ reasoning abilities are varied and dynamic. It is concluded that chemistry reasoning is multidimensional and should be assessed through integrated, data-driven methods. The study recommends the adoption of formative, analytics-supported assessments to inform differentiated instruction and promote deeper conceptual and ethical engagement in chemistry education.]]>
Copyrights © 2025
