The impact of chatgpt-assisted problem-based learning on students physics achievement and their chatgpt acceptance

As a new technology, ChatGPT has been integrated into teaching and learning, but there is still a lack of studies on ChatGPT-assisted problem-based learning in physics. Therefore, this study investigates its integration in relation to students’ academic achievement and ChatGPT acceptance. This study used a quasi-experimental design with 65 students who divided into an experimental class and a control class. Data were collected via pre- and post-tests and analyzed using Rasch analysis, ANCOVA, t-test, and DIF, supported by validated instruments and expert review. The study found that students in the experimental group, who used ChatGPT-assisted Problem-Based Learning (PBL), showed significantly higher post-test scores in physics compared to the control group. Wright Map and DIF analyses revealed stronger performance on higher-order conceptual items among the experimental group. ANCOVA confirmed a large effect size (η² = 0.715) for the intervention. Rasch analysis of ChatGPT acceptance showed that the experimental group had higher agreement with most items, and independent t-tests indicated significantly greater acceptance across variables like attitude, habit, and behavioral intention. These findings support ChatGPT-assisted PBL as effective for both learning outcomes and technology acceptance.