ABSTRACT While research on students’ difficulties in explaining evolution has shifted focus from isolated concepts to the role of intuitive ideas, few studies have yet applied cognitive linguistics. Drawing on frame semantics and conceptual blending, this study examines how upper-secondary students use semantic frames to make sense of natural selection. It employs a qualitative analysis of semantic frames in 31 upper-secondary students’ written explanations of natural selection before and after interacting with an educational visualisation of antibiotic resistance. The results show that students used semantic frames such as adjusting, immunisation, or learning in their explanations. Furthermore, students frequently maintained their initial framing, even after instruction, sometimes leading to mixed and hybrid explanations that combined intuitive and scientific ideas. Conceptual blending was applied to explain how such mixed explanations may emerge when elements from different semantic frames and mental spaces are fused. The findings suggest that semantic frame analysis of student explanations may complement traditional concept-based assessments of natural selection understanding. The results also imply that evolution teaching may require explicit attention to reframing students’ interpretations, thus helping them to transition from everyday conceptions to scientific understandings. The article concludes with implications for visualisation designs for teaching evolution.

ABSTRACT In response to growing concerns about low participation in science, technology, engineering, and mathematics (STEM) careers, there is a need for more understanding of the underlying cognitive and motivational mechanisms affecting students’ STEM career expectations. This study examined whether relative strengths of abilities and attitudes in science, mathematics, and Chinese language could predict STEM career expectations among junior secondary school students. Data were from the Hong Kong PISA 2022 sample. Data were analysed using person-centred approaches and structural equation modelling (SEM). Results revealed six distinct attitude classes (Passionate, Prefer Chinese, Prefer Science/Math, Prefer Chinese/Math, Persistent, and Uncommitted) and four ability profiles (High-achieving, Above-average, Below-average, and Low-achieving). Students in the Prefer Science/Math class with a clear science/mathematics preference over the Chinese language were the most likely to select STEM careers, whereas the other groups (e.g. Passionate, Prefer Chinese, Persistent, and Uncommitted) were less likely to pursue them. Additionally, the Prefer Science/Math class positively mediated the relationships between student ability profiles and career expectations, whereas the Prefer Chinese/Math class negatively mediated these relationships. Overall, this study underscores the importance of attitudes toward science and mathematics relative to other subjects (e.g. language) in shaping adolescents’ career expectations and future STEM participation.

ABSTRACT Our understanding of elementary teachers’ science subject matter knowledge (SMK) development through teaching experience is incomplete. In this study, we explored the relationships with factors that may be associated with SMK. 346 elementary teachers completed one test for evolution and another test for force. They provided information about formal opportunities (college science coursework, professional development programmes), features of teaching experience (years, required curriculum materials, weekly team interactions), and personal characteristics (interest in science, science identity). The strongest predictors of teachers’ science SMK were personal characteristics of interest and identity. Teaching more grade levels and being required to use curriculum materials were negatively associated with SMK for evolution. Interacting with colleagues at least once a week and years of experience were not related to SMK. Similarly, formal opportunities like college science coursework and hours in science professional development did not predict SMK.

ABSTRACT Textual descriptions and simplified examples (non-formal explanations) are commonly used to introduce science topics to students, e.g. in geoscience, but may be ambiguous and understood in different ways, potentially leading to misconceptions. This exploratory study considers how image schemata, mental patterns based on bodily experiences, can be used to investigate how non-formal explanations can be (mis)understood. We illustrate this approach by analysing undergraduate textbooks' non-formal introductions to isostasy, a geoscience concept incorporating the physics concepts of elasticity and buoyancy. Our analysis reveals that the explanations can be understood through different image schemata, some of which are not well aligned to the current scientific understanding of isostasy. Moreover, many of these image schemata can logically lead to misconceptions and do not accurately reflect the corresponding physics. The analysis also highlights how introductions may require that students reconceptualise isostasy to transition to the current scientific understanding. Image schemata may therefore be useful for teachers and education researchers to identify productive, and avoid problematic, understandings when using non-formal explanations.

ABSTRACT As immersive technologies gain traction in education, understanding their psychological impacts becomes crucial. This study developed and validated the Motivation toward Virtual Reality Scale and examined its associations with science anxiety, academic buoyancy, and grit. Three studies were conducted with middle school students. Study I (N = 301) involved exploratory factor analysis for scale development. Study II (N = 292) confirmed construct validity and reliability using confirmatory factor analysis, internal consistency metrics (α, ω, λ), Item Response Theory (IRT), measurement invariance, and criterion-related validity. Study III (N = 418) tested a structural model examining the mediating roles of academic buoyancy and grit between VR motivation and science anxiety. Results supported a 14-item, three-factor structure with strong psychometric properties. IRT analyses showed high item discrimination, and invariance analyses confirmed stability across gender. VR motivation was positively associated with VR attitude and acceptance. Structural equation modelling revealed that academic buoyancy and grit mediated the link between VR motivation and science anxiety. Grounded in Self-Determination Theory and related frameworks, findings highlight the role of VR motivation in reducing science anxiety and enhancing students’ resilience in technology-integrated learning environments.

ABSTRACT This study aims to investigate upper secondary school students’ engagement in computer-supported collaborative learning while interacting with a digital interactive visualisation of the carbon cycle. Task-based interviews were conducted with student pairs who used the visualisation to complete a carbon cycle activity, focusing on the transformation and movement of carbon compounds between carbon reservoirs. Drawing on an activity theory framework, the analysis identified contradictions and alignments among elements of the learning activity system. Findings indicate that students’ ability to construct system-based explanations was enhanced by access to key scientific concepts, effective interaction with the visualisation, and guided instructional support. Students’ development of explanations was sometimes obstructed when they engaged task-irrelevant concepts, lacked access to concepts or made unhelpful interpretations of the visualised information. Furthermore, a collaborative and socially supportive learning environment played a critical role in fostering collective sense-making and active engagement with the complexities of Earth system processes.

ABSTRACT Learners’ conceptions of scientists are often assumed to influence their academic interests – but empirical evidence for this relationship remains limited and inconsistent. Moreover, little is known about how students perceive scientists beyond STEM domains and how these conceptions relate to their academic interests. These blind spots are uncovered in the present study by investigating secondary school students’ conceptions of educational scientists compared to natural scientists, and how these conceptions relate to their study interest and self-image. We conducted semi-structured interviews with 64 students and used Epistemic Network Analysis to analyse their conceptions. Findings reveal that natural scientists are more strongly associated with scientific activities and cognitive abilities than educational scientists. Notably, students with high study interest and self-image in educational sciences have more inaccurate conceptions, conflating educational scientists with pedagogical roles. Across disciplines, students with low interest and self-image tend to hold pessimistic, stereotypical views of scientists, whereas those with high interest and self-image view them more positively, associating them with communion-oriented activities and prosocial characteristics. This study highlights the need to promote more accurate student views of different scientific disciplines to better align their expectations and career choices, ultimately improving their engagement and reducing dropout rates.

ABSTRACT This study investigates the capability of ChatGPT-4o to generate high-quality inquiry-based science lesson plans, that is, aligning all elements of a written lesson plan to students’ learning about procedural and epistemic aspects of science instead of gaining subject matter knowledge. Using an exploratory sequential mixed methods design, we analysed N = 60 biology lesson plans generated by the research team across four key topics (cell biology, genetics/evolution, human biology, ecology) and five scientific inquiry practices (microscopy, observing, experimenting, modelling, reflecting the nature of science) from the German curriculum . First, lesson plans were quantitatively evaluated using a modified version of Großmann and Krügers’ [(2024). Assessing the quality of science teachers' lesson plans: Evaluation and application of a novel instrument. Science Education, 108(1), 153–189. https://doi.org/10.1002/sce.v108.1] scoring rubric, which assessed ten quality criteria with substantial interrater agreement (Cohen's κ = .67). Second, based on the score distribution, we conducted qualitative analyses to identify strengths and weaknesses in ChatGPT-generated inquiry-based lesson plans. Results revealed considerable variation in lesson plan quality, with n = 22 lesson plans achieving less than 50% of the maximum possible score and only n = 5 lesson plans reaching 75% or higher. While the lesson plans demonstrated particular strengths in content accuracy and learning outcome alignment, they exhibited significant weaknesses in addressing students’ inquiry-related conceptions and maintaining consistent focus on scientific inquiry. While ChatGPT-4o successfully generated some high-quality lessons plans for inquiry-based instruction, many lesson plans shifted inappropriately towards subject matter knowledge acquisition rather than scientific inquiry processes. This discrepancy between our initial prompts and the resulting lesson plans indicates that while large language models like ChatGPT-4o demonstrate potential as preliminary planning tools, they necessitate thorough evaluation and substantive modification by teachers possessing robust pedagogical content knowledge. The findings emphasise that the advent of generative artificial intelligence does not diminish the importance of professional knowledge in teacher education but rather transforms how this knowledge is applied in lesson planning processes.

ABSTRACT This article explores the potential of generative artificial intelligence (GAI) tools to create adaptive learning pathways for inquiry-based learning (IBL) science instruction. The research compares tailored pathways from ChatGPT 4.0 and Llama 3.1 for differentiated inquiry instruction in science subjects. A qualitative approach was employed using a case study design to provide thick descriptions of Pathways from ChatGPT 4.0 and Llama3.1in supporting teachers to cater to personalised IBL instruction. Five science teachers were purposefully selected to generate and implement adaptive learning pathways for IBL instruction. Content analysis of GAI-generated pathways and teacher semi-structured interviews revealed insights into the performance of the GAI models and the usefulness of the pathways in teaching inquiry lessons. The findings revealed that LLM-based GAI engines could be used to address the individual learning needs of students in IBL science classrooms. Adaptive learning pathways showed great potential in differentiating instruction. ChatGPT 4.0 was seen to be more robust when compared with Llama 3.1. Challenges observed included difficulty keeping track of differentiated assessments, time management, ethical concerns and AI biases. These findings have implications for exploiting GAI tools when crafting IBL instruction. Some recommendations for teacher practice and research in developing adaptive learning pathways are also discussed.