Assessing students' understanding of computational modeling in physics

Abstract

Computational modeling is increasingly included in secondary physics curricula. In classroom practice, however, students often work with pre-designed computational models rather than constructing or modifying models themselves. Research on modeling in science education indicates that meaningful engagement with models requires meta-modeling knowledge: an understanding of models' nature, purpose, use, and limitations in scientific inquiry. Yet little is known about how secondary students reason about computational models and how this meta-modeling knowledge can be systematically analyzed. This study adapts the Framework for Modeling Competence (FMC) to the context of physics computational modeling in order to examine students’ meta-modeling knowledge. The resulting Framework for Computational Modeling Competence (FCMC) characterizes students’ reasoning about computational modeling across five epistemic aspects (Nature, Multiple, Purpose, Testing, and Changing), each articulated across three levels of understanding. Data were collected through semi-structured interviews with 36 upper secondary pre-university physics students in the Netherlands. Students were asked to reason about two computational physics models, and their utterances were analyzed using the FCMC's aspect–level combinations. The analysis shows that students most frequently demonstrated meta-modeling knowledge related to the Purpose and Testing aspects of computational models. In contrast, reasoning corresponding to the highest level of understanding was absent for the aspects Nature, Multiple, and Changing. These findings suggest that while students can use computational models to interpret or compare results, they experience greater difficulty reasoning about computational models as epistemic tools involving assumptions, alternative representations, and model revision. The study provides an analytical framework for examining students’ meta-modeling knowledge in computational modeling and highlights the need for instructional approaches that explicitly support epistemic reasoning about computational models in physics education.