Multimodal Large Language Models for Diagnostic Feedback Analytics in STEM Learning Platforms

“Creative Pragmatics for Active Learning in STEM Education” opens with a chapter that introduces Creative Pragmatics as a flexible and evolving approach aimed at fostering active learning in STEM. This introductory chapter, provided as open access, advocates for competence-based education that prepares learners for the unpredictable and multidimensional challenges of today’s world. It challenges the traditional view of scientific knowledge as stable and complete, urging educators and students to develop competence and skills to navigate the unpredictable challenges of today’s science and society. Spanning 13 chapters, the book features contributions from a diverse array of scholars and practitioners, from Stanford, Columbia, and the University of Nebraska in the USA, with experts from Sweden, Germany and several Danish universities and university colleges. These contributors explore themes such as creative learning strategies, dynamic teacher-student interactions, innovative assessment methods, and the design of learning environments. Each chapter further develops the idea of Creative Pragmatics, emphasising a holistic approach that fosters creativity, interdisciplinary collaboration, and active student engagement. This collection is a valuable resource for educators, researchers, and policymakers who seek to rethink STEM education. The book offers practical insights and a ‘bigger picture’ perspective for fostering educational environments that enable students to develop the competencies needed to engage with complexity. It encourages an exploration of the world that nurtures curiosity, critical thinking, and creative agency, thereby promoting professional and civic responsibility towards environmental, societal, technological, and scientific challenges. Bringing together a transdisciplinary and global team with professional backgrounds spanning Asia, the Americas, and Europe, this book is part of a series on European Science Education Research and Practice, contributing significantly to the field.

Science, Technology, Engineering, and Mathematics (STEM) education curriculum and instructional design continuously undergo reforms that aim to prepare learners for the challenges of the 21st century (Hoeg & Bencze, 2017; Pietarinen et al., 2017). In particular, STEM education has adopted strategies that integrate modern technologies in teaching and learning to enhance knowledge construction and application among learners and societies. In some countries, STEM education reforms are fuelled by socio-economic and political imperatives that seek to promote social justice (Mnguni, 2018). More recently, the emergence of the Covid-19 pandemic has forced schools and universities to adopt online teaching methods to reduce the coronavirus's spread. Consequently, researchers are exploring strategies for the incorporation of online teaching and learning methods. The effectiveness of these strategies and their impact on the students' conceptualization of STEM knowledge, its application, and relevance are continuously being investigated.

Pneumothorax requires rapid recognition and accurate interpretation of chest X-rays (CXRs), particularly in acute settings where delays can have serious consequences. With the emergence of advanced image interpretation models capable of visual analysis, their diagnostic reliability in radiology practice remains to be determined. This study aimed to assess the diagnostic performance of three state-of-the-art systems in detecting pneumothorax using a large, well-annotated dataset. A total of 10,675 CXRs from the publicly available SIIM-ACR Pneumothorax Segmentation dataset were analyzed. Three multimodal models (GPT-4o, Gemini 2 Pro, and Claude 4 Sonnet) were evaluated using a uniform, image-based approach. Each model's binary outputs (presence: 1, absence: 0) were compared with reference results to determine accuracy, sensitivity, specificity, precision, and F1 scores. Additional subgroup analyses were conducted across pneumothorax size categories: small, medium, and large. Pairwise statistical comparisons were performed using McNemar's test. Sensitivity, specificity, and overall accuracy are reported with corresponding 95% confidence intervals. The prevalence of pneumothorax in the dataset was 22.3% (n = 2,379). All models demonstrated high specificity (above 0.90) but consistently low sensitivity (0.16-0.36). The best overall performance was observed with Gemini 2, which achieved an accuracy of 0.79 and specificity of 0.95, whereas Claude 4 showed greater sensitivity (0.20-0.34) across lesion-size categories. Diagnostic performance improved with increasing pneumothorax size, but smaller lesions remained difficult to identify. Pairwise comparisons confirmed statistically significant differences among all evaluated systems (P < 0.050). In this large-scale evaluation, the tested models exhibited strong reliability in identifying normal examinations but limited ability to detect subtle or small pneumothoraxes. Despite high specificity, low sensitivity limits the use of current Multimodal large language models as rule-out tools for pneumothorax. With continued refinement, these models may eventually support radiologists by improving workflow efficiency and diagnostic confidence. Automated systems capable of high specificity but low sensitivity should not be relied upon to exclude pneumothorax. However, they may serve as valuable assistants for confirming positive findings and prioritizing urgent cases in busy clinical workflows.

The aim of this paper is to examine the use of digital technologies in Science, Technology, Engineering, and Mathematics Education (STEM). We will discuss the effectiveness of the teaching-learning process in terms of the elements that could possibly promote learning with the use of Information and Communication Technologies (ICT) in STEM education. This will be done, first, by taking a learner-centred approach to an activity that students carried out using ICT to performing a task set by the teacher (cognitive engagement in the task, motivation, nature of knowledge built). The aim was to understand how ICT could be a cognitive aid for the student. Second, a teacher-centred perspective to the development of prescribed tasks (form, knowledge carried by the task) was used to identify how ICT can be adapted to aid student learning in STEM education.

This article provides a bibliometric overview of publications on eLearning trends in STE(A)M teaching and learning to give readers a better understanding of the current state of research in the field. The main objective of this study is to provide bibliometric data on publications on online teaching and learning trends for science, technology, engineering, and mathematics education (STEM) teaching and learning purposes printed in journals included in the Scopus database in the years 2011-2023. For the bibliometric analysis, STEM learning, STEM teaching, online education, bibliometric review keywords were used, and 136 documents from the Scopus database were chosen. The collected data of the publications scanned and published in the parameters of the study were subjected to a bibliometric analysis based on seven categories: number of articles and citations per year, most influential countries, most prolific author, most prominent affiliations, funding institutions, publication source, and subject areas. Network diagrams and bibliometric analyses were created using the Scopus database analysis. Most of the articles were published between 2016 and 2022. The United States of America, the United Kingdom, and China were among the top-three most productive countries, and the United States of America produced the most publications. The number of citations to publications indexed in the Scopus database is growing steadily and reached its peak in 2022 (178 citations). The most prolific author on this subject is Minichiello, A., with four publications. In addition, Stanford University and Utah State University have maximum publishing partners. By funding 16 publications for online STEM teaching and learning, the National Science Foundation has shown leadership. The topic areas of the publications’ distribution were looked at. The articles’ respective fields of study were social sciences and computer science. This study offers a vision for future research as well as a worldwide view of online learning for STEM teaching and learning.

Rural, regional and remote (RRR) communities and industries in Australia cannot currently produce or attract the workforce needed to survive, making skills and qualifications in science, technology, engineering and mathematics (STEM) increasingly important. Yet student engagement in STEM education in RRR schools remains low, with limited numbers of young people either moving into further STEM education post-secondary, or accessing readily available STEM-related jobs in RRR areas. Currently many rural children and young people are not exposed to, nor recognize the diverse ways in which STEM knowledge is required and used in their world. We propose that if young people are to increasingly engage with STEM and continue onto STEM-related careers, they must be able to see connections between their “school” learning of STEM and the STEM knowledge that is enacted in rural work and life. We also suggest that for this to change, there should be increased visibility of “place-based” knowledges, including Aboriginal STEM knowledges, in RRR communities to promote enhanced student engagement with STEM. In this paper we explore these ideas by drawing on Foucault and Bourdieu understandings to develop a methodological framework – thePlace-based STEM- alignment Frameworkfor the purposes of exposing alternate STEM knowledges. We argue that the nuanced and critical methodological approach applied in the development of thePlace-based STEM-alignment Framework, is necessary in order to generate this analytical tool and provide data that will allow us the scope to “reset” current understandings of STEM knowledges. The framework design provides us with the methodological vehicle to identify possible reasons for the invisibility of STEM knowledge and practices in the local fabric of RRR communities and to examine enablers and/or barriers to engagement in STEM learning. The framework must be a practical tool for use in the field, one that can be used in RRR communities to engage, children and young people, in STEM, in a way that is meaningful and that aligns with their everyday experience of RRR life. Finally, the framework has to work to enable alternative perspectives to be exposed that will advance methodological considerations of STEM.

Online tutoring has been found to have a positive impact on student learning in STEM education. The aim of this study was to explore how and to what extent students supported their own and other students' processes of inquiry during online tutoring. More specifically, question and answer sessions (Q&As) were studied, and online tutoring transcripts were analyzed using the Relationship of Inquiry coding scheme adapted from the Community of Inquiry framework. The students interacted voluntarily with each other and with the teachers. The results show that the students supported both their own process of inquiry as well as other students' process of inquiry. Furthermore, the results indicate that students acquired metacognitive development through self- and co-regulation when they expressed teaching presence. Overall, we conclude that Q&As have the potential to support student learning when working on assignments in STEM education.

This Study aim to investigate the effects of integrated inquiry-based and STEM education learning management on Thai grade 8 students&amp;rsquo; analytical thinking and learning achievement. The research employed a one-group experimental design with 37 grade 8 students from a public school in Khon Kaen Province, Thailand, selected through purposive sampling. The instruments used in this study included an inquiry-based STEM education integrated learning management plan, a learning achievement test on motion and force, and an analytical thinking test. Data were collected through pretests and posttests on learning achievement and analytical thinking, and ongoing assessments during the learning activities. The data were analyzed using percentage, mean score, standard deviation, effectiveness index (E1/E2), and a paired samples t-test. The results indicated that the overall effectiveness of the learning management, calculated as the E1/E2 ratio, was 77.20/77.57, meeting the predetermined criteria of 75/75. Furthermore, significant increases in analytical thinking and learning achievement were observed at the statistical level of 0.5. These findings support the effectiveness of integrating inquiry-based learning with STEM education in enhancing students&amp;rsquo; analytical thinking and learning achievement, contributing valuable insights to educational practices in science education.

This study aims to determine teachers' knowledge level about AI-driven learning platforms in STEM education, to determine the extent to which language teachers apply AI-driven learning platforms in STEM education, and to see whether the level of knowledge, attitudes and readiness of teachers is balanced with its application. This study uses a quantitative and qualitative design or mixed method and applies a correlation approach. The participants of this study involved 33 English teachers in Indonesia. The results of this study conclude that the level of knowledge of language teachers on AI-driven platforms in STEM learning is still classified as intermediate. The application of AI-driven platforms in STEM learning for English teachers is also in the intermediate category. Language teachers have not yet entered the high level of application of AI-driven platforms in STEM learning, although a small number of teachers can reach this level. The correlation between teacher knowledge of AI-driven platforms in STEM learning is directly proportional to their application in learning practices. However, the data shows no correlation between teacher attitudes and readiness towards AI-driven platforms in STEM learning and the application of AI-driven platforms in STEM learning.

STEM Education intends to have Integrated Learning in the disciplines of Science, Technology, Mathematics and Mathematics to produce critical, innovative and creative students who contribute to the country's development. STEM education has a role not only as a channel to process and produce local skilled workers, even as a leader in the country's development economy. However, STEM education scenarios in Malaysia are dealing with various issues such as the lack of STEM teaching materials, the reduction in the number of students who continue study in STEM-related fields in the middle and tertiary rankings and low achievements in international level markers studies such as the Program for International Student Assessment (PISA ) and Trends in International Mathematics and Science Study (TIMSS). These issues are related to one of the main challenges, namely STEM learning / teacher teaching quality that is not interesting and consistent and teacher quality in the STEM field. Increasing the quality of STEM learning, the quality of teacher teaching and the quality of teachers needs to be done as one of the proactive steps to strengthen STEM Education. If this effort is not carried out, the dictatorship of the country becomes incompetent in the global ranking and lags behind. This concept paper naturally discusses the issues and challenges of teacher factors in the education of STEM education and so on several reserves and initiatives to overcome the STEM learning challenges / teacher teaching quality that is not interesting and consistent and teacher quality is raised.

Integrated STEM education has the potential to develop important skills, competencies and dispositions needed in the twenty-first century. Children’s foundational knowledge is shaped in the early years and primary teachers have therefore been recognised as the gatekeepers to STEM education. Many teachers, however, feel unprepared to teach integrated STEM. This study responds to the call for more evidence on the types of learning experiences needed to prepare STEM-literate preservice teachers. Using pre-/post intervention data from surveys and STEM task analysis documents, as well as reflective journal entries, we report on 30 preservice teachers’ initial understandings of integrated STEM education and explore their evolution across a 12-week intervention, discussing key learning experiences that reportedly led to these changes. The findings suggest that preservice teachers had deeper, more critical understandings of STEM education as a result of the intervention and began to develop a pedagogical toolkit to support STEM learning in the classroom. These developments can be attributed to the powerful interplay that existed between participants’ experiences as STEM learners in coursework and as STEM teachers during field practice. In an effort, however, to address low initial confidence in engineering, we inadvertently tipped the scales of disciplinary balance, at the expense of mathematics.

Students of modern times are learning fast and passionately. They love the interactivity and are very interested in science. Robotics can be safely assigned as engineering for students. This is the first step toward disciplines such as programming, mathematics, science, and algorithms. Thus, the objective of the study is to investigate the feasibility of implementing robotics in our educational system to develop learning in STEM subjects. Therefore, the main purpose of this study was to investigate the effect of mBot robot activities in STEM subjects as part of project-based learning, to determine their satisfaction and perceptions of the project carried out. The methodology used for this research is the descriptive study model, which combines qualitative and quantitative research methods. The selected sample of 14 students was a purposive sample, since the participants were the students who had implemented the robotics competition project. Quantitative data is collected from questionnaires, while qualitative data is generated from open-ended survey questions. From the results of this study, it has been shown that the integration of the mBot robot into the educational system will undoubtedly increase the effectiveness of teaching and learning STEM education. Also, it was found that the level of students' creativity, attitude, and motivation increased with the inclusion of the mBot robot during project-based learning in STEM.

Although inquiry-based learning originated in physics and science education, where students are involved in authentic scientific practices, it also occurs in mathematical or technological contexts. This approach to learning mirrors the procedure and thinking that scientists, engineers, and innovators use in the real world. As a result, inquiry-based learning naturally suits to related disciplines such as science, technology, engineering and mathematics or STEM disciplines. In this paper we present examples of STEM related projects designed for students in order to solve real-life problems using digital technologies to support inquiry represented by modelling or experimental approach. Examples of interdisciplinary projects involve motion of sportsmen, sky or scuba divers, or solving other problems using various programming platforms. Students working on interdisciplinary projects effectively use and deepen their knowledge from different subjects and develop skills to use smart technologies and coding. As a result, they move from the role of passive users to active learners and designers.

This article explores a case study that presents the integration of field trips within an intensive delivery mode to enhance preservice teachers’ engagement and learning in STEM education. Using a design thinking and structured learning cycle approach, the collaboration between Victoria University and Wyndham Tech School offered applied learning opportunities through tutorials, field trips, and assessments embedded in VU’s Block Model ® . The impact of this collaborative framework is presented with qualitative research of 35 preservice teachers’ perceptions of their learning through focus groups and reflective assessment submissions. Thematic analysis identified five key themes, exploring both benefits and challenges of the unit, including enriching the student experience, the role of design thinking in learning, the transformative impact of the field trip, perceptions of future teaching practice, and engagement barriers. This case study discusses curriculum design, unit structure, pedagogy, and the impact of the program on student experience, while addressing logistical challenges associated with intensive learning programs.

Human-GenAI interaction for active learning in STEM education: State-of-the-art and future directions