Models are at the core of scientific reasoning and science education. They are especially crucial in scientific and educational contexts where the primary objects of study are unobservables. While empirical science education researchers apply philosophical arguments in their discussions of models and modeling, we in turn look at exemplary empirical studies through the lense of philosophy of science. The studied cases tend to identify modeling with representation, while simultaneously approaching models as tools. We argue that such a dual approach is inconsistent, and suggest considering models as epistemic artifacts instead. The artifactual approach offers many epistemic benefits. The access to unobservable target systems becomes less mysterious when models are not approached as more or less accurate representations, but rather as tools constructed to answer theoretical and empirical questions. Such a question-oriented approach contributes to a more consistent theoretical understanding of modeling and interpretation of the results of empirical research.
11 June 2025 in Journal of Science Education and Technology
Integrating digital technologies in science education requires innovative methods to bridge theoretical knowledge and practical classroom application. This study examines a seminar at an Austrian university that was redesigned to address pre-service chemistry teachers’ professional knowledge regarding technology implementation in high school classes via a service-learning approach. By combining questionnaires with formative evaluation through reflective learning diaries, our study captures the multifaceted nature of professional knowledge acquisition. Central to this work is the use of natural language processing within a Computational Grounded Theory framework to analyze students’ reflective diaries. We employed techniques such as word embeddings and topic modeling to extract latent themes and patterns in student texts. We revealed that, despite mixed results from conventional self-report measures and deductive qualitative content analysis, students’ written reflections offered rich insights into their learning processes when we investigated them through the lens of natural language processing. Along the diary entries, our analysis uncovered shifts in emphasis from, e.g., broad cultural perceptions of teaching and learning to more detailed considerations of lesson planning and technology integration. These nuanced insights underscore the complementary value of natural language processing in identifying underlying patterns of reflective practice that traditional assessments may overlook. Although the study is limited by its small sample size and methodological constraints, the findings suggest that incorporating computational techniques can enhance the formative assessment of free writing in teacher education programs. Overall, the results motivate us to advocate for the integration of computational text analysis as a promising tool in evaluating and fostering the complex interplay of professional knowledge and reflective practice in technology-enhanced science education.
19 September 2022 in Journal of Chemical Education
Systematic analyses of repeated measures of cognitive load during realistic, complex learning tasks are still lacking. Previous findings are based on simulations and the use of a set of multiple short problem-solving tasks. For the present study, lower secondary level pupils ( N = 360) worked on three worked examples on the topic of acids in their chemistry lessons. Cognitive load (invested mental effort and perceived task difficulty), interest, and self-assessed understanding were measured at seven points during the completion of each example. The study examines differences between repeated measures of cognitive load and interest and the relationship between the variables. The results show that the values of the repeated measures differ significantly and are poorly represented by the mean value across the measurement points. In addition, the pairwise relationship between the variables (invested effort, perceived task difficulty, interest, and self-assessed understanding) is not constant at different measurement points. Further analyses also show that different intervention conditions during the learning phase and prior knowledge are additional influencing factors.
For centuries, chemists have researched acid-base reactions and developing numerous different acid-base concepts. Consequentially, it is difficult to teach about acid-base reactions in a clear and conceptually coherent manner. We aim at contributing to teaching about acid-base chemistry by an intervention introducing upper secondary school chemistry students to an acid-base concept that is scientifically appropriate and learner-centred. This modernized version of the Brønsted-Lowry acid-base concept appears sufficiently complex for the target group. Additionally, it emphasizes the donor-acceptor concept and the reaction mechanism. In order to evaluate the gain in declarative knowledge during our intervention, we developed and piloted a multiple-choice knowledge test with Austrian upper secondary students (N = 136). By estimating a Rasch model (WLE-Rel. = 0.63, Infitmin = 0.90, Infitmax = 1.02, Outfitmin = 0.89, Outfitmax = 1.00, d = −0.44–1.8), we constructed a pre-test (11 MC items, 6 open-ended questions) and a post-test (11 MC items, 6 open-ended questions) based on key concepts of the topic for our intervention study (N = 57). Overall, the test proved a reliable tool to estimate learners’ declarative knowledge about acid-base reactions in the course of the intervention study, showing a significant knowledge increase.