Computers & Education最新文献

{"title":"Exploring the prospects of multimodal large language models for Automated Emotion Recognition in education: Insights from Gemini","authors":"Shuzhen Yu ,&nbsp;Alexey Androsov ,&nbsp;Hanbing Yan","doi":"10.1016/j.compedu.2025.105307","DOIUrl":"10.1016/j.compedu.2025.105307","url":null,"abstract":"<div><div>Emotions play a pivotal role in daily judgments and decision-making, particularly in educational settings, where understanding and responding to learners’ emotions is essential for personalized learning. While there has been growing interest in emotion recognition, traditional methods, such as manual observations and self-reports, are often subjective and time-consuming. The rise of AI has led to the development of Automated Emotion Recognition (AER), offering transformative opportunities for educational reform by enabling personalized learning through emotional insights. However, AER continues to face challenges, including reliance on large-scale labeled databases, limited flexibility, and inadequate adaptation to diverse educational contexts. Recent advancements in AI, particularly Multimodal Large Language Models (MLLMs), show promise in addressing these challenges, though their application in AER remains underexplored. This study aimed to fill this gap by systematically evaluating the performance of Gemini, a pioneering MLLM, in image-based AER tasks across five databases: CK+, FER-2013, RAF-DB, OL-SFED and DAiSEE. The analysis examined recognition accuracy, error patterns, emotion inference mechanisms, and the impact of image preprocessing techniques — such as face cropping, bilinear interpolation, and super-resolution — on the model’s performance. The results revealed that Gemini achieved high emotion recognition accuracy, especially in distinguishing emotional polarities across all databases. Image preprocessing significantly improved the recognition of basic emotions, though its effect on academic emotion recognition was minor. The confusion in academic emotion recognition stemmed from Gemini’s limited understanding of academic emotion features and its insufficient ability to capture contextual cues. Building on the results, this study outlines specific future research directions from both technological and educational perspectives. These findings offer valuable insights for advancing MLLMs in educational applications.</div></div>","PeriodicalId":10568,"journal":{"name":"Computers & Education","volume":"232 ","pages":"Article 105307"},"PeriodicalIF":8.9,"publicationDate":"2025-03-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143768892","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"教育学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Meaning particles and waves in MOOC video lectures: A transpositional grammar guided observational analysis","authors":"Christopher Hughes","doi":"10.1016/j.compedu.2025.105308","DOIUrl":"10.1016/j.compedu.2025.105308","url":null,"abstract":"<div><div>This study employs transpositional grammar to analyze the structure of meaning in MOOC video lectures. An observation guide was developed using meaning forms in transpositional grammar – speech, sound, body, object, image, and text – as the superordinate categories and populated with sub-categories informed by a literature review, pilot study, and consultation with a panel of experts. Through a detailed examination of 25 video lectures, meaning forms are categorized as \"particles\" (discrete units) and \"waves\" (extended patterns). Transpositional grammar reveals how particles, like conversational cues, enhance accessibility, while waves, such as consistent narration and branding, create structural coherence and reinforce institutional presence. Findings highlight how transpositional grammar can elucidate recurring patterns in meaning-making, offering a framework for designing educational media that maximizes cognitive engagement in online learning contexts.</div></div>","PeriodicalId":10568,"journal":{"name":"Computers & Education","volume":"236 ","pages":"Article 105308"},"PeriodicalIF":8.9,"publicationDate":"2025-03-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144169087","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"教育学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Development and implementation of a generative artificial intelligence-enhanced simulation to enhance problem-solving skills for pre-service teachers","authors":"Jieun Lim ,&nbsp;Unggi Lee ,&nbsp;Junbo Koh ,&nbsp;Yeil Jeong ,&nbsp;Yunseo Lee ,&nbsp;Gyuri Byun ,&nbsp;Haewon Jung ,&nbsp;Yoonsun Jang ,&nbsp;Sanghyeok Lee ,&nbsp;Jewoong Moon","doi":"10.1016/j.compedu.2025.105306","DOIUrl":"10.1016/j.compedu.2025.105306","url":null,"abstract":"<div><div>Effective teachers should be equipped to solve complex problems across diverse instructional and learning contexts. However, many teacher training programs struggle to bridge the gap between theoretical knowledge to real-world applications. The current study tackles this challenge by developing a generative artificial intelligence <strong>(</strong>GenAI)-enhanced simulation to improve preservice teachers’ problem-solving abilities. Using design-based research (DBR), we created a virtual environment that integrates problem-based learning (PBL) with GenAI technology. The simulation was rigorously refined through expert review and usability testing before being implemented in a teacher training program. We evaluated its effectiveness by comparing three groups: (1) a text-based scenario, (2) a rule-based simulation, and (3) a GenAI-enhanced simulation. Pre- and post-test results showed significant improvements in problem-solving skills for both the rule-based and GenAI-enhanced simulation groups compared to the text-based scenario group. Notably, qualitative findings revealed that students reported heightened realism and immersion in the GenAI-enhanced simulation, attributing this to more dynamic interactions with AI agents that helped them better contextualize PBL and increased their motivation. Our study findings contribute design principles for developing GenAI-enhanced simulations in teacher education, offering promising insights into leveraging AI technology to create more engaging and effective training experiences.</div></div>","PeriodicalId":10568,"journal":{"name":"Computers & Education","volume":"232 ","pages":"Article 105306"},"PeriodicalIF":8.9,"publicationDate":"2025-03-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143735069","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"教育学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Modeling student teachers’ self-regulated learning of complex professional knowledge: A sequential and clustering analysis with think-aloud protocols","authors":"Lingyun Huang ,&nbsp;Ying Zhan ,&nbsp;Shen Ba","doi":"10.1016/j.compedu.2025.105310","DOIUrl":"10.1016/j.compedu.2025.105310","url":null,"abstract":"<div><div>There has been much discussion regarding the positive relationship between self-regulated learning (SRL) and technological pedagogical content knowledge (TPACK) development for student teachers. This study continued this claim and adopted advanced analytical methods to explain how SRL influences TPACK learning. Think-aloud protocols from 39 participants were collected and transcribed when they were learning TPACK by designing technology-infused lessons with nBrowser, a computer-based learning environment. Based on models, nine critical SRL events were retrieved from participants‘ think-aloud protocols and analyzed through sequential clustering analysis. The results show two SRL groups indicating distinct self-regulatory sequential patterns. One group had a shorter sequence length and dominantly enacted elaboration activities (Low-SRL group), while the other had longer sequence lengths and engaged in diverse SRL activities (High-regulation group). Relating to TPACK performance indicated by the quality of lesson plans, the results reveal that the participants in the High-SRL group outperformed their counterparts in the Low-SRL group. The findings are consistent with previous evidence and provide implications for practitioners about the importance of student teachers’ self-regulation trajectories.</div></div>","PeriodicalId":10568,"journal":{"name":"Computers & Education","volume":"233 ","pages":"Article 105310"},"PeriodicalIF":8.9,"publicationDate":"2025-03-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143845283","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"教育学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Relationships between interactive network patterns and students’ cognitive and emotional processes: Evidence from the individual and group levels","authors":"Zhongjian Liu ,&nbsp;Hainuo Liu ,&nbsp;Yuan Tian ,&nbsp;Xinxiao Nie ,&nbsp;Mengjie Liu","doi":"10.1016/j.compedu.2025.105309","DOIUrl":"10.1016/j.compedu.2025.105309","url":null,"abstract":"<div><div>Analysing the processes of effective computer-supported collaborative learning (CSCL) is a prominent area of research in education. Within the community of enquiry framework, group social interaction patterns are essential for enhancing learners' cognitive and emotional processes. This study examined the online collaborative learning processes of 66 university students over an eight-week blended course. Process and perception data were collected from multiple sources, including surveys, log records, and sentence content analysis. Employing social network and content analyses, we examined the network patterns formed by different group interactions and identified individual and group differences. The multilevel analysis results indicated that group density and centralisation were generally not correlated with the learners’ expressions of cognitive viewpoints, perceived task value, or satisfaction. However, the individual degree centrality was significantly and positively correlated with the expression of viewpoints, perceived task value, and satisfaction, whereas the betweenness centrality was negatively correlated with satisfaction. Furthermore, the perceived task value mediates the relationship between the degree of centrality and satisfaction. This study expands the literature on the CSCL process research by incorporating social network parameters at different levels, shedding light on the relationships between individuals and groups in collaboration, and offering insights into the design and instructional practices of online collaborative learning.</div></div>","PeriodicalId":10568,"journal":{"name":"Computers & Education","volume":"232 ","pages":"Article 105309"},"PeriodicalIF":8.9,"publicationDate":"2025-03-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143759706","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"教育学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Effects of integrating self-regulation scaffolding supported by chatbot and online collaborative reflection on students’ learning in an artificial intelligence course","authors":"Chia-Wen Tsai ,&nbsp;Lynne Lee ,&nbsp;Michael Yu-Ching Lin ,&nbsp;Yih-Ping Cheng ,&nbsp;Chih-Hsien Lin ,&nbsp;Meng-Chuan Tsai","doi":"10.1016/j.compedu.2025.105305","DOIUrl":"10.1016/j.compedu.2025.105305","url":null,"abstract":"<div><div>Artificial intelligence (AI) includes complex concepts and could be difficult for non-computer or information students, and they may experience difficulties in an AI course. In order to develop students' AI skills, regulate their academic stress, and reduce student loneliness, the researchers in this study integrated self-regulation scaffolding (supported by a chatbot designed in this study) with online collaborative reflection, and investigated their effects on students' learning. The researchers conducted a quasi-experiment to explore the effects of self-regulation scaffolding and online collaborative reflection. The participants in this experiment were 116 undergraduates from three classes (groups) of a non-computer department taking a compulsory course titled ‘Introduction to Artificial Intelligence’. The experimental groups in this study included the first class (G1) simultaneously receiving the interventions of self-regulation scaffolding and online collaborative reflection, as well as the second class (G2) only receiving the intervention of self-regulation scaffolding. The last class (G3), that received a traditional teaching method (non-self-regulation scaffolding and non-online collaborative reflection) in an online AI course, served as the control group. According to the statistical analysis in this study, the self-regulation scaffolding approach in G2 significantly promoted participants' AI skills and fostered their ability to regulate academic stress compared to the control group. However, G1 students who received concurrent online collaborative reflection did not demonstrate the expected effects of enhancing their learning compared with those (G2) who did not receive it. This study represents an early attempt to implement self-regulation scaffolding and online collaborative reflection integrated with chatbot in an online AI course. The results of this study could serve as a reference for online course designers, educators, and scholars planning to adopt these teaching methods, especially for those focusing on implementing online AI education and educational technologies (such as chatBot).</div></div>","PeriodicalId":10568,"journal":{"name":"Computers & Education","volume":"232 ","pages":"Article 105305"},"PeriodicalIF":8.9,"publicationDate":"2025-03-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143680692","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"教育学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Self-regulated learning and technological integration competency: A multilevel latent profile analysis of preschool teachers","authors":"Jing Li ,&nbsp;Barry Bai","doi":"10.1016/j.compedu.2025.105295","DOIUrl":"10.1016/j.compedu.2025.105295","url":null,"abstract":"<div><div>Existing research has independently examined teachers' beliefs, emotions, and knowledge related to technology use, but few studies have explored how these components can be combined to form technological integration competency (TIC). Furthermore, the role of self-regulated learning (SRL) in predicting TIC remains underexplored. This study employed multilevel latent profile analysis (MLPA) to identify TIC profiles at both teacher and school levels and examineed their associations with continuous usage intention and technology-enabled productivity. Using data from 3256 teachers nested within 274 schools, single-level latent profile analysis revealed three teacher profiles: Novice, Competent, and Advanced. Teachers in more advanced profiles exhibited higher usage intention and reduced productivity. At the school level, three profiles, Developing, Progressing, and Leading, were identified based on the distribution of teacher profiles, with substantial differences in school-level usage intention and productivity. SRL emerged as a significant predictor of TIC profiles at both levels, with higher SRL increasing the likelihood of teachers and schools belonging to better profiles. The discussion emphasizes integrating beliefs, emotions, and knowledge into teacher training programs to promote TIC while considering the Chinese cultural context.</div></div>","PeriodicalId":10568,"journal":{"name":"Computers & Education","volume":"232 ","pages":"Article 105295"},"PeriodicalIF":8.9,"publicationDate":"2025-03-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143672849","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"教育学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Design and evaluation of a repository of toolkits based on design thinking for socio-educational intervention with vulnerable migrant women","authors":"M. P. (María del Pino) Díaz-Pereira ,&nbsp;J. (Joseba) Delgado-Parada ,&nbsp;M.-C. (María-Carmen) Ricoy ,&nbsp;M. J. (Manuel José) Fernández-Iglesias","doi":"10.1016/j.compedu.2025.105294","DOIUrl":"10.1016/j.compedu.2025.105294","url":null,"abstract":"<div><div>Digital resources are valuable in socio-educational interventions with vulnerable migrant women for developing creative skills. Few existing repositories offer systematized resources for social educators to apply creative methodologies, such as design thinking. In the framework of a European project, this paper aims to design and evaluate Thinklusion, a repository of toolkits based on design thinking for socio-educational interventions with vulnerable migrant women. Before designing the repository (structure and content), vulnerable migrant women's profile was identified. Based on the ADDIE model, a needs study was carried out through interviews with four representatives of European institutions. To evaluate didactic and technical aspects of Thinklusion, 144 professionals completed an online questionnaire based on the 71362:2020 standard. In addition, Wave software was applied to the repository to analyze its web accessibility. The results and conclusions draw a vulnerable migrant women's profile as a starting point for designing inspiring models and therefore digital toolkits, incorporated in the repository. In the didactic evaluation, the contents were deemed respectful, clear, and applicable. It is suggested to favor autonomous learning and motivation. In the technical evaluation, the quality of the textual and iconic content was highlighted, as well as the presence of Accessible Rich Internet Applications attributes. In this dimension, it is necessary to provide self-help tools and solve color contrasts between texts and backgrounds. No significant differences were identified in the evaluation according to gender. The next phases of the research will be aimed at solving the errors, training educators, and implementing the toolkits in socio-educational practice.</div></div>","PeriodicalId":10568,"journal":{"name":"Computers & Education","volume":"231 ","pages":"Article 105294"},"PeriodicalIF":8.9,"publicationDate":"2025-03-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143629432","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"教育学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Is it externally stimulated or internally driven? A study on the influencing mechanisms of Chinese graduate students' intention to continue online learning","authors":"Honghong Cai ,&nbsp;Mingjie Pan ,&nbsp;Leilei Huang","doi":"10.1016/j.compedu.2025.105293","DOIUrl":"10.1016/j.compedu.2025.105293","url":null,"abstract":"<div><div>The COVID-19 pandemic has significantly propelled the progression of online instruction in China. Existing studies have demonstrated that the quality of teachers' online teaching, as well as students' self-regulated learning (SRL), has a considerable impact on undergraduates' intention to persist with online learning. However, few studies have simultaneously considered these two aspects within a unified analytical framework, particularly regarding their influence on graduate students' intention to continue online learning. Therefore, this study aims to identify teachers' e-readiness and graduate students' SRL that enable students' online learning outcomes and, ultimately, their intention to continue online learning. Data were collected via a survey of 10,346 graduate students from 40 top universities in China, who had participated in online course learning. The results derived from the structural equation models and conditional process models indicated that teachers' e-readiness significantly influenced online learning outcomes, which in turn predicted students' intention to continue online learning. Additionally, it was discovered that self-regulated learning negatively moderated both the direct and indirect relationships between teachers' e-readiness and graduate students' intention to continue online learning. The constructed equations representing graduate students' intention to continue online learning suggested that relative to teachers' e-readiness, graduate students' SRL exerted a greater influence on their intention to continue online learning, with the two factors exhibiting substitution effects under specific circumstances.</div></div>","PeriodicalId":10568,"journal":{"name":"Computers & Education","volume":"231 ","pages":"Article 105293"},"PeriodicalIF":8.9,"publicationDate":"2025-03-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143627567","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"教育学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Does exposure to open educational practices foster open thinking? Exploring direct and indirect effects and other influential factors","authors":"Insung Jung ,&nbsp;Jihyun Lee","doi":"10.1016/j.compedu.2025.105283","DOIUrl":"10.1016/j.compedu.2025.105283","url":null,"abstract":"<div><div>The growing adoption of open educational practices (OEPs) in higher education calls for empirical research to understand their impact on learners, particularly in fostering open thinking as a key learning outcome. Moreover, the roles of individual traits, course design, and cultural factors in shaping open thinking remain underexplored, especially across different national contexts. This study aims to address these gaps by examining the direct and indirect effects of exposure to OEPs on open thinking and other influential factors. Data were collected from 667 participants across four countries (South Korea, Japan, the U.S., and the Philippines). Exposure to OEPs, individual factors (e.g., demographics, extroverted personality, technology acceptance attitudes), course factors (e.g., learner-centered activities, support, control), cultural factors (e.g., background, multicultural experiences), and open thinking were measured. Hierarchical multiple regression, moderation and mediation analyses, and structural equation modeling (SEM) were employed to explore direct, indirect, and moderating effects, whereas multi-group SEM was used to assess cross-national differences in the relationships between these variables. The findings highlight the significant impact of OEP exposure on fostering open thinking, with the technology acceptance attitude serving as a crucial mediating factor. While individual factors such as having extroverted personality, multicultural experience, and technology acceptance have shown limited direct or indirect influence, learner-centered course design has emerged as a key factor in promoting open thinking. Cultural differences across South Korea, Japan, the U.S., and the Philippines were evident, indicating that the relationships among these factors vary by context, shaping the development of open thinking in distinct ways.</div></div>","PeriodicalId":10568,"journal":{"name":"Computers & Education","volume":"238 ","pages":"Article 105283"},"PeriodicalIF":8.9,"publicationDate":"2025-03-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144714550","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"教育学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}