Answers are grounded in this paper and cite their sources.
Single-group intervention study demonstrates improved achievement and confidence in high school students, indicating that interpretable AI modeling aids complex chemistry learning.
This study examined the multidimensional effects of a decision tree-based scientific modeling program on high school students’ learning of redox reactions and explored the applicability of findings from prior small-scale studies in a larger classroom context. Participants were 96 students from a general high school, and the program was implemented over ten instructional sessions using a single-group pre-test–post-test design. The instructional program integrated scientific modeling practices within the 5E learning cycle. Using the Orange3 platform, students externalized their judgment rules as data through model construction, compared their judgments with AI predictions through model testing, and reconstructed concepts by diagnosing errors through model revision. Program effects were examined in terms of academic achievement, understanding of redox reaction models, judgment confidence, learning experiences, and perceived potential for transfer. Results showed statistically significant improvements in overall academic achievement from pre-test to post-test (p < 0.001), with a very large effect size (d = 2.43) across all model domains. Students’ confidence in redox judgment also increased substantially (d = 0.93), alongside the development of a more pluralistic view of models that recognized contextual appropriateness and limitations. In addition, the interpretable AI-based modeling process functioned as cognitive scaffolding that supported metacognitive reflection during model construction, testing, and revision. Notably, a large proportion of students reported that decision tree-based reasoning could be applied beyond chemistry learning to everyday decision making and future academic or career contexts, indicating features of high-road transfer. These findings suggest that interpretable AI-supported scientific modeling practice can foster robust cognitive, affective, and process-oriented learning outcomes in regular high school classrooms.
No takes yet. Share an insight, caveat, or question.
Moon et al. (2026) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: