Computers & Education最新文献

{"title":"Webcam use and its role in children's engagement and achievement during extended remote learning","authors":"Sharlyn Ferguson-Johnson ,&nbsp;Allison M. Ryan ,&nbsp;Kai S. Cortina","doi":"10.1016/j.compedu.2025.105443","DOIUrl":"10.1016/j.compedu.2025.105443","url":null,"abstract":"<div><div>Online classes substitute in-person instruction in K-12 general education, yet whether children's webcam use in online classes is meaningful for their learning remains poorly understood. This study examined children's webcam use and academic outcomes during the COVID-19 pandemic using data from 1426 elementary students (49% male, 47% female, 93% White) across 65 classrooms, exploring how individual- and classroom-level webcam use, along with socioeconomic and demographic factors and remote learning duration, explained variation in childrens' engagement and achievement. In May 2021, students reported their remote learning routines and online class participation since May 2020; responses were then linked with 2020–2021 MSTEP achievement scores and demographic data. Correlational analyses revealed individual webcam use varied as a function of children's perceived engagement, academic achievement, socioeconomic disadvantage, and gender identity, while classrooms differed in mean webcam use based on socioeconomic and demographic composition, remote learning duration, and mean engagement and achievement. MANOVA results indicated classroom webcam use norms had a larger effect on engagement during remote learning relative to individual use. Specifically, children in classrooms with <em>universal webcam-on norms</em> reported significantly higher mean emotional and behavioral engagement compared to children in classrooms with <em>moderate/mixed</em> or <em>low webcam-on norms.</em> Multilevel models revealed classmates' mean webcam use explained a significant portion of the variance in children's 2020–2021 achievement and was a stronger predictor of achievement than individual webcam use. Overall, findings underscore the need to consider classroom-level webcam use policies and incentives when developing and refining effective online classroom structure and curricula for elementary school student populations.</div></div>","PeriodicalId":10568,"journal":{"name":"Computers & Education","volume":"241 ","pages":"Article 105443"},"PeriodicalIF":10.5,"publicationDate":"2026-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144900521","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 AI-based personalized recommender system to improve students’ self-regulated learning and academic performance","authors":"Xinyi Luo,&nbsp;Sikai Wang,&nbsp;Khe Foon Hew","doi":"10.1016/j.compedu.2025.105486","DOIUrl":"10.1016/j.compedu.2025.105486","url":null,"abstract":"<div><div>Students' ability to manage their own learning is crucial for academic success, but many struggle with self-regulation, often leading to disengagement. Past studies have explored tactics to boost students' self-regulated learning (SRL), such as writing self-reflective reports and incorporating prompts in video lectures. These tactics, however, often lack timely, personalized feedback due to the labor-intensive nature of such support. In this study, we report the development and implementation of an innovative large language model (LLM)-based recommender system named <em>SRLAdvisor</em> to address these limitations. The system combines a web interface with a series of LLM-based agents, allowing students to receive immediate, personalized feedback. By processing students' natural language interactions in real-time, <em>SRLAdvisor</em> dynamically profiles their self-regulation and delivers personalized recommendations. This study involved 22 postgraduate students and employed a sequential explanatory mixed-methods approach to evaluate students’ self-regulation and learning performance. We found near-perfect agreement between the system detection and human coding. <em>SRLAdvisor</em> exhibited greater recommendation precision in self-monitoring and self-evaluation tasks. Students who perceived the system as highly useful showed significantly greater self-reported SRL gains than those with lower perceptions. Additionally, a <em>k-means</em> cluster analysis indicated that students who engaged more frequently with <em>SRLAdvisor</em> achieved significantly greater improvements in learning performance. These findings underscore the potential of leveraging LLMs to deliver personalized recommendations. However, they also reveal potential concerns, such as over-reliance on AI and prompt monotony, that need to be further examined. These issues are particularly relevant in the context of LLM-supported SRL interventions and warrant further discussion.</div></div>","PeriodicalId":10568,"journal":{"name":"Computers & Education","volume":"241 ","pages":"Article 105486"},"PeriodicalIF":10.5,"publicationDate":"2026-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145315150","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":"Team collaborative norms enhance students’ behavioral engagement in team-based flipped classrooms: A multilevel analysis","authors":"Hui-Min Lai ,&nbsp;Pi-Jung Hsieh ,&nbsp;Fang-Chuan Ou Yang","doi":"10.1016/j.compedu.2025.105468","DOIUrl":"10.1016/j.compedu.2025.105468","url":null,"abstract":"<div><div>The flipped classroom (FC) model replaces traditional in-class lectures with student-centered, team-based activities. In the FC model, in-class sessions are typically conducted in teams to promote active participation, group discussion, problem-solving, and collaborative learning. However, a review of previous studies reveals that the factors contributing to behavioral engagement in team-based FCs remain underexplored. Thus, we aim to explore team-based FCs by integrating the theories of expectancy–value and social motivation. Given that collaborative norms are crucial for enhancing students’ behavioral engagement in teams, we propose that team collaborative norms moderate the relationship between individuals’ beliefs (value and competence beliefs) and behavioral engagement. We collected data from 808 students across 238 teams in 28 team-based FCs and used a hierarchical linear model to examine the hypothesized relationships. The results indicated a positive relationship between utility value and students’ behavioral engagement, as well as between interest value and engagement in team-based FCs. Conversely, there was a negative relationship between learning anxiety and students’ behavioral engagement. No significant relationship was found between attainment value and engagement. Additionally, self-efficacy and team collaborative norms were positively associated with students’ behavioral engagement in team-based FCs. Notably, the moderating effect of team collaborative norms on the relationships between utility value and behavioral engagement and between interest value and behavioral engagement was stronger under high team collaborative norms than under low ones. This study contributes to existing research by validating the role of team collaborative norms in students’ behavioral engagement within team-based FCs.</div></div>","PeriodicalId":10568,"journal":{"name":"Computers & Education","volume":"241 ","pages":"Article 105468"},"PeriodicalIF":10.5,"publicationDate":"2026-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145261664","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":"A generative artificial intelligence-enhanced multiagent approach to empowering collaborative problem solving across different learning domains","authors":"Lanqin Zheng,&nbsp;Zhe Shi,&nbsp;Lei Gao","doi":"10.1016/j.compedu.2025.105489","DOIUrl":"10.1016/j.compedu.2025.105489","url":null,"abstract":"<div><div>Collaborative problem-solving skills are among the most important skills in the 21st century. However, learners exhibit significant deficiencies in terms of their collaborative problem-solving skills. Emerging artificial intelligence (AI) technologies have given rise to transformative opportunities to facilitate collaborative problem solving through the introduction of adaptive learning mechanisms in educational settings. This study proposes a generative artificial intelligence (GenAI)-enhanced multiagent approach that aims to promote collaborative problem solving across different learning domains. The study also examines the effectiveness of this GenAI-enhanced multiagent approach to collaborative problem solving. In total, 234 college students participated in two empirical studies that focused on different tasks but had the same purpose and procedure. Experimental group 1 engaged in collaborative problem-solving with the assistance of a GenAI-enhanced multiagent approach. Experimental group 2 engaged in collaborative problem solving via a chatbot-based approach. The control group engaged in traditional collaborative problem-solving without any support. Data were collected through the pretest, posttest, and collaborative problem-solving process records and interview records. Both quantitative and qualitative methods were employed to analyze the data. The results indicated that compared with chatbot-based and traditional approaches, the GenAI-enhanced multiagent approach had more significant effects on learning achievements, knowledge elaboration, and collaborative problem-solving performance and skills. The implications of these findings are discussed in depth with the goal of advancing the use of GenAI to empower collaborative problem solving.</div></div>","PeriodicalId":10568,"journal":{"name":"Computers & Education","volume":"241 ","pages":"Article 105489"},"PeriodicalIF":10.5,"publicationDate":"2026-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145359263","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":"Adaptive vs. planned metacognitive scaffolding for computational thinking: Evidence from generative AI-supported programming in elementary education","authors":"Jinfang Liu ,&nbsp;Yi Zhang ,&nbsp;Wei Li ,&nbsp;Qiyun Wang ,&nbsp;Pingxiu Niu ,&nbsp;Xue Zhang","doi":"10.1016/j.compedu.2025.105473","DOIUrl":"10.1016/j.compedu.2025.105473","url":null,"abstract":"<div><div>Metacognitive scaffolding plays a pivotal role in supporting the development of computational thinking (CT) among elementary students. Compared to planned scaffolding, adaptive scaffolding supported by generative AI (GAI) agents offers a promising option by providing real–time, adaptive and personalized response. This study investigated how adaptive metacognitive scaffolding (AMS) and planned metacognitive scaffolding (PMS) influenced elementary students’ CT in a GAI–supported programming environment. 74 students were assigned to AMS, PMS, or control conditions. Using a mixed-methods approach, we examined CT task performance, CT skill use, metacognitive regulation, and cognitive load through standardized assessments, programming artifacts, discourse logs, video-coded behaviors, classroom observation records, cognitive load ratings. Results showed that AMS significantly enhanced contextualized CT performance and promoted more complex, diverse skill use compared to PMS and control groups. AMS also fostered more recursive metacognitive behaviors and reduced mental efforts, while PMS supported higher meaningful engagement. These findings highlight the potential of adaptive scaffolding powered by GAI to support CT development in elementary programming education.</div></div>","PeriodicalId":10568,"journal":{"name":"Computers & Education","volume":"241 ","pages":"Article 105473"},"PeriodicalIF":10.5,"publicationDate":"2026-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145236446","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":"Ex-Edchat: Historic retrospective of X/Twitter #Edchat","authors":"K. Bret Staudt Willet ,&nbsp;Jeffrey P. Carpenter ,&nbsp;Hunhui Na","doi":"10.1016/j.compedu.2025.105476","DOIUrl":"10.1016/j.compedu.2025.105476","url":null,"abstract":"<div><div>For more than a decade, education-related X/Twitter hashtags facilitated networking and resource-sharing among educators with related interests and needs, resulting in self-reported impacts on practice. #Edchat was one of the first such hashtags and attracted substantial attention as an affinity space for educators. This study retrospectively explores long-term and large-scale digital trace X/Twitter data associated with #Edchat from October 2008 to May 2023, analyzing more than 15 million tweets in terms of changes in volume (e.g., daily tweets) and content (e.g., questions, replies, hyperlinks, co-occurring hashtags, language features). Findings suggest that #Edchat’s initial success led to impressive growth, followed by change in the nature of content and a long period of steady decline. Specific social dynamics associated with the hashtag, such as the decline of its associated synchronous chat, as well as technical factors (e.g., platform updates, policy changes) appear to have influenced #Edchat’s volume and content. Quantifying the shifting nature of this long-standing affinity space contributes to understanding the opportunities and challenges educators may encounter on social media broadly and highlights the importance of supporting and developing educators’ digital literacy.</div></div>","PeriodicalId":10568,"journal":{"name":"Computers & Education","volume":"241 ","pages":"Article 105476"},"PeriodicalIF":10.5,"publicationDate":"2026-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145359261","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":"Incremental project-based robot programming: Effects on young children's computational thinking, executive function, and learning behavioral patterns","authors":"Yuchen Chen ,&nbsp;Xinli Zhang ,&nbsp;Lailin Hu ,&nbsp;Gwo-Jen Hwang ,&nbsp;Haoran Xie ,&nbsp;Yun-Fang Tu","doi":"10.1016/j.compedu.2025.105471","DOIUrl":"10.1016/j.compedu.2025.105471","url":null,"abstract":"<div><div>Robot programming is an age-appropriate means to cultivate children's computational thinking (CT) and executive function (EF). Project-based learning (PBL) is commonly applied to teach robot programming. PBL involves complex, large-scale projects consisting of multiple sub-projects. Each sub-project introduces new knowledge but offers few opportunities to connect it with previously learned knowledge, thus misaligning with children's cognitive development. In comparison, incremental PBL (IPBL) is a tailored design of PBL that begins with basic knowledge, with each new sub-project building on all learned knowledge from previous sub-projects and progressively adding new knowledge to achieve complete project learning. Hence, this research integrated IPBL into robot programming to develop an incremental project-based robot programming (I-PBRP) approach, and evaluated its effect on young children's CT, EF, and learning behavioral patterns. Ninety-five children aged 5–6 engaged in 12-week interventions. They were randomly assigned to the I-PBRP group and the conventional project-based robot programming (C-PBRP) group. Results manifested that the I-PBRP group achieved better performance than the C-PBRP group in CT and in the three components of EF (inhibition, working memory, and cognitive flexibility) over time. The progressive behavioral analysis manifested that the I-PBRP approach promoted superior performance in robot programming activities and more positive learning behaviors for children. This research has implications for robot programming teaching approaches for young children's CT and EF development.</div></div>","PeriodicalId":10568,"journal":{"name":"Computers & Education","volume":"241 ","pages":"Article 105471"},"PeriodicalIF":10.5,"publicationDate":"2026-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145315169","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":"Mapping student-AI interaction dynamics in multi-agent learning environments: Supporting personalized learning and reducing performance gaps","authors":"Zhanxin Hao ,&nbsp;Jie Cao ,&nbsp;Ruimiao Li ,&nbsp;Jifan Yu ,&nbsp;Zhiyuan Liu ,&nbsp;Yu Zhang","doi":"10.1016/j.compedu.2025.105472","DOIUrl":"10.1016/j.compedu.2025.105472","url":null,"abstract":"<div><div>Multi-agent AI systems, which simulate diverse instructional roles such as teachers and peers, offer new possibilities for personalized and interactive learning. Yet, the specific patterns of student-AI interaction and their pedagogical implications remain unclear. This study explores how university students engaged with multiple AI agents during a six-module course, and how these interactions influenced cognitive outcomes (learning gains) and non-cognitive factors (motivation, technology acceptance). Through the analysis of dialogue data, two core engagement patterns were identified: co-construction of knowledge and co-regulation. Students with lower prior knowledge relied more on co-construction of knowledge sequences and showed higher learning gains and post-course motivation. In contrast, students with higher prior knowledge engaged more in co-regulation behaviors but demonstrated limited learning improvement. Technology acceptance increased across all groups. These findings suggest that multi-agent systems can effectively support differentiated engagement and help reduce performance gaps by adapting to students' varying needs. This study makes three innovative contributions to existing research: it is based on a long-term formal curriculum, moving beyond fragmented, short-term interactions; it investigates a multi-agent collaborative mechanism that simulates diverse pedagogical roles (e.g., AI teacher, AI teaching assistance, AI classmates), distinguishing this work from single-agent systems; and it explores differentiated interaction modes based on prior knowledge, providing critical teaching implications for personalized learning.</div></div>","PeriodicalId":10568,"journal":{"name":"Computers & Education","volume":"241 ","pages":"Article 105472"},"PeriodicalIF":10.5,"publicationDate":"2026-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145261910","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":"Unpacking self-regulation and social interaction in “Study With Me” videos through large-scale analytics","authors":"Tongxi Liu ,&nbsp;Liping Deng ,&nbsp;Yujie Zhou","doi":"10.1016/j.compedu.2025.105488","DOIUrl":"10.1016/j.compedu.2025.105488","url":null,"abstract":"<div><div>As social media platforms are increasingly used to facilitate informal learning, “Study With Me” (SWM) videos have garnered substantial popularity. Despite their widespread use, empirical research on these videos remains in its infancy. This study investigates the characteristics of SWM videos, providing a comprehensive understanding of their affordances for self-regulated learning and social interaction. Specifically, advanced machine learning techniques were applied to analyze 393 SWM videos and 164,611 associated comments on YouTube. A modified topic modeling approach identified emerging themes and patterns in the comment data, while sentiment analysis assessed emotional tone and examined how specific video features influenced users' self-regulation. The analysis revealed that comments primarily focused on SWM video features, self-regulation, and social interaction. Positive sentiment appeared in about half of the comments, praising elements such as ambient music and visual aesthetics for enhancing emotional engagement and motivation. Various features of SWM videos, such as lighting, music, and in-video text, support learners’ self-regulation across motivational, emotional, and social dimensions. This study highlights the potential of social media as a versatile educational tool and encourages stakeholders to leverage such platforms to expand and enrich learning opportunities.</div></div>","PeriodicalId":10568,"journal":{"name":"Computers & Education","volume":"241 ","pages":"Article 105488"},"PeriodicalIF":10.5,"publicationDate":"2026-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145315151","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-regulation plus individual interests? A design-based study on the development of a GenAI-empowered platform for self-directed out-of-class reading","authors":"Mengru Pan ,&nbsp;Chun Lai ,&nbsp;Kai Guo ,&nbsp;Jing Huang","doi":"10.1016/j.compedu.2025.105474","DOIUrl":"10.1016/j.compedu.2025.105474","url":null,"abstract":"<div><div>Self-directed out-of-class reading contributes to language and general learning, yet many learners struggle with it and need support. Generative Artificial Intelligence (GenAI) holds the potential to provide personalized support that can facilitate self-directed out-of-class reading. This study adopted a design-based research (DBR) methodology to unravel the effective design of GenAI-empowered self-directed out-of-class reading by comparing different ways of integrating a ChatGPT chatbot into an online reading platform in supporting the provision of reading texts and delivering self-regulated learning (SRL) support. The DBR study included three cycles: a conventional approach in Cycle 0 (i.e., a preset pool of reading materials plus human-delivered SRL training), a chatbot-empowered personalized SRL platform in Cycle 1 (i.e., GenAI-recommended reading materials and chatbot-supported personalized SRL training), and a chatbot-empowered interest-based reading with personalized SRL platform in Cycle 2 (i.e., GenAI-recommended interest-based reading materials and chatbot-supported personalized SRL training). This DBR study lasted three semesters, involving one hundred and sixteen English as a foreign language (EFL) students from three classes at a university in China. Qualitative data from open-ended surveys and interviews after each cycle were used to refine the platform design in the subsequent cycle. Quantitative data, including pre- and post-surveys on self-regulated reading strategy use and self-directed reading, along with log data, were analyzed to compare the effects of the platform design on students' self-regulated strategy use and self-directed reading across three cycles. The results indicated that the chatbot-empowered interest-based reading with personalized SRL training showed the greatest potential in enhancing students’ self-regulated reading strategy use and self-directed reading. This study contributes to the existing literature by elucidating the design principles that enhance self-directed out-of-class reading.</div></div>","PeriodicalId":10568,"journal":{"name":"Computers & Education","volume":"241 ","pages":"Article 105474"},"PeriodicalIF":10.5,"publicationDate":"2026-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145261874","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}