Generative AI-supported problem-based learning for enhancing computational thinking and chemistry achievement in engineering chemistry

Abstract

The increasing adoption of Generative Artificial Intelligence (GenAI) in higher education has created new opportunities to enhance active learning; however, empirical evidence regarding its integration with Problem-Based Learning (PBL) in Engineering Chemistry remains limited, particularly for simultaneously fostering Computational Thinking (CT) and chemistry achievement. This study investigates the effectiveness of integrating GenAI with PBL in an undergraduate Engineering Chemistry course within a Mechanical Engineering Education program. In the intervention, GPT-4 was embedded as an AI learning scaffold that guided students through Socratic questioning, provided formative feedback, and supported problem decomposition and analysis during collaborative PBL activities rather than supplying direct answers. Employing a quasi-experimental pretest-posttest control group design, the study involved 68 undergraduate students (n = 34 experimental; n = 34 control) selected through cluster random sampling. Data were collected using three validated instruments: a Computational Thinking Test (CTT, 30 items, ? = .87), a Chemistry Achievement Test (CAT, 35 items, ? = .84), and a Learning Observation Sheet (LOS, 20 items, ICC = .91). Data were analyzed using descriptive statistics, Shapiro-Wilk normality tests, Levene's homogeneity tests, independent-samples t-tests, ANCOVA, and Cohen's d effect size analysis. The results revealed significant improvements in both CT skills (t(66) = 8.47, p .001) and chemistry achievement (t(66) = 9.12, p .001) in the experimental group compared with the control group. ANCOVA controlling for pretest scores confirmed significant treatment effects for CT (F(1,65) = 127.84, p .001) and chemistry achievement (F(1,65) = 109.07, p .001), with large practical effects (Cohen's d = 2.14 and 2.06, respectively). These findings provide empirical support for integrating AI-supported scaffolding within structured PBL environments to strengthen computational thinking and conceptual understanding in Engineering Chemistry, while extending the pedagogical application of AI-enhanced constructivist learning in engineering education.