AI-Integrated problem-based learning and critical thinking in elementary science: an exploratory study

Abstract

This study aims to determine the association between Artificial Intelligence (AI)-integrated Problem-Based Learning (PBL) and critical thinking skills of fifth-grade students. The research employed a single-group experimental design with a sample size of 28 students. Data collection was conducted through pre-tests and post-tests designed to measure students' critical thinking skills regarding physical science concepts. Data analysis included normalized gain (N-Gain) scores to determine improvement levels, a Paired Sample t-Test to evaluate instructional effectiveness, and Cohen’s (dz) to measure the effect size of the instructional model. The results showed that the students' average pre-test score was 54.8, which increased to 79.0 in the post-test (mean difference = 24.2). The average N-Gain score was 0.54 (moderate category), indicating a significant improvement in critical thinking skills. The Paired Sample t-Test yielded a t-value of t(27) = 13.8 (p < 0.001), confirming that the increase in post-test scores was statistically significant. Furthermore, the effect size analysis resulted in a value of 2.63 (large category), demonstrating that AI-integrated PBL may have a substantial association with student performance. The findings suggest that integrating AI into PBL may strengthen skills in analysis, evaluation, and inference.