Exploring physics students’ attitudes toward ChatGPT using the ABC model

The integration of artificial intelligence tools such as ChatGPT into educational settings has sparked a paradigm shift in higher education, necessitating a deeper understanding of students’ attitudes toward these technologies. The ABC model, which delineates attitudes into affective, behavioral, and cognitive components, provides a robust framework for such investigations. Prior studies have applied this model broadly across multiple disciplines. However, little is known about its applicability in physics education, where a strong emphasis on analytical reasoning and quantitative problem-solving might influence attitudes uniquely. Addressing this gap, we conducted a cross-sectional survey study using an online questionnaire administered to N = 1,189 physics students enrolled at German universities. We developed an instrument, adapted from prior research, to assess students’ attitudes toward ChatGPT in the context of physics learning. The validity of the instrument’s hypothesized three-factor structure was then evaluated via confirmatory factor analysis. The results paint a clear picture: The three-factor solution demonstrated satisfactory global fit (CFI = 0.95, RMSEA = 0.05, SRMR = 0.04) and significantly outperformed alternative two- and one-factor models based on likelihood ratio tests and information criteria. The results thus affirm the empirical validity of this instrument in capturing physics students’ attitudes toward ChatGPT according to the ABC model, contributing to a nuanced understanding of learner perspectives on ChatGPT in discipline-specific educational contexts. Additionally, an overview is provided of physics students’ attitudes toward learning with ChatGPT by analyzing their responses on the item level. Implications for educational practice and future research are discussed.