Pub Date : 2026-01-24DOI: 10.1016/j.compedu.2026.105579
T.-C. Chang, Alice R.P. Li, C.-Y. Wang Ph.D, John J.H. Lin Ph. D
Analyzing the complex dialogue central to computer-supported collaborative learning is crucial for understanding learning processes, yet remains a significant challenge for educational researchers due to the labor-intensive nature of manual coding and the semantic limitations of traditional computational methods. Recent advancements have highlighted the potential of Large Language Models (LLMs) to move beyond mere automation, demonstrating an ability for inference without task-specific data that is characteristic of artificial general intelligence. To harness this potential, this study introduced and evaluated a human-AI collaborative framework (CGT-LLM) that integrates LLMs into computational grounded theory. Specifically, CGT-LLM focuses on learning analytics for rich discursive data. Applied to dialogue from a climate change collaborative simulation game, the framework was evaluated against a supervised bidirectional encoder representations from transformers (BERT) baseline. The performance of the framework approached human expert-level performance in categories related to explicit instructions, numerical data, or direct statements of intent crucial to game objectives, while also demonstrating promising capability in identifying more abstract and less obvious themes. The findings demonstrate that the researcher's role in computational grounded theory remains critical, particularly in exploring data diversity during the discovery phase, and making final interpretive judgments for abstract themes during the classification phase. This framework thus positions LLMs as a valuable assistant rather than as a replacement for human expertise, providing educators and researchers with a tool to gain deeper, more scalable insights into collaborative learning processes, and offering potential to inform the design of timely pedagogical interventions.
{"title":"From Automation to Thinking: The Role of AGI in Discourse Analysis of Computer-Supported Collaborative Learning Based on Computational Grounded Theory","authors":"T.-C. Chang, Alice R.P. Li, C.-Y. Wang Ph.D, John J.H. Lin Ph. D","doi":"10.1016/j.compedu.2026.105579","DOIUrl":"https://doi.org/10.1016/j.compedu.2026.105579","url":null,"abstract":"Analyzing the complex dialogue central to computer-supported collaborative learning is crucial for understanding learning processes, yet remains a significant challenge for educational researchers due to the labor-intensive nature of manual coding and the semantic limitations of traditional computational methods. Recent advancements have highlighted the potential of Large Language Models (LLMs) to move beyond mere automation, demonstrating an ability for inference without task-specific data that is characteristic of artificial general intelligence. To harness this potential, this study introduced and evaluated a human-AI collaborative framework (CGT-LLM) that integrates LLMs into computational grounded theory. Specifically, CGT-LLM focuses on learning analytics for rich discursive data. Applied to dialogue from a climate change collaborative simulation game, the framework was evaluated against a supervised bidirectional encoder representations from transformers (BERT) baseline. The performance of the framework approached human expert-level performance in categories related to explicit instructions, numerical data, or direct statements of intent crucial to game objectives, while also demonstrating promising capability in identifying more abstract and less obvious themes. The findings demonstrate that the researcher's role in computational grounded theory remains critical, particularly in exploring data diversity during the discovery phase, and making final interpretive judgments for abstract themes during the classification phase. This framework thus positions LLMs as a valuable assistant rather than as a replacement for human expertise, providing educators and researchers with a tool to gain deeper, more scalable insights into collaborative learning processes, and offering potential to inform the design of timely pedagogical interventions.","PeriodicalId":10568,"journal":{"name":"Computers & Education","volume":"85 1","pages":""},"PeriodicalIF":12.0,"publicationDate":"2026-01-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146047986","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}
Pub Date : 2026-01-20DOI: 10.1016/j.compedu.2026.105577
Kozlova Zoya, Cera Raphael, Hoyer Christoph, Flegr Salome, Kuhn Jochen, I. Hofer Sarah
{"title":"The Impact of the Testing Environment on Gender Differences in Mental Rotation Performance: Does Virtual Reality Make a Difference?","authors":"Kozlova Zoya, Cera Raphael, Hoyer Christoph, Flegr Salome, Kuhn Jochen, I. Hofer Sarah","doi":"10.1016/j.compedu.2026.105577","DOIUrl":"https://doi.org/10.1016/j.compedu.2026.105577","url":null,"abstract":"","PeriodicalId":10568,"journal":{"name":"Computers & Education","volume":"33 1","pages":""},"PeriodicalIF":12.0,"publicationDate":"2026-01-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146014859","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}
Pub Date : 2026-01-19DOI: 10.1016/j.compedu.2026.105578
Xinheng Song , Yue Zhang , Zhaolin Lu , Linci Xu , Hengheng Shen
With the widespread integration of generative AI tools into educational contexts, understanding their influence on learners’ cognitive and emotional processes has become increasingly critical. While AI holds potential for enhancing creativity, its double-edged impact on neurocognitive and emotional processes still requires further investigation. This study investigates the impact of generative AI-based learning tools on the creative thinking learning process. Participants were divided into two groups: a generative AI design group and a traditional design group. They completed tasks employing the divergent brainstorming creative method and the structured innovation TRIZ method. During these tasks, both facial expressions and functional near-infrared spectroscopy (fNIRS) data were collected to explore the effects of generative AI-assisted creative thinking education on students’ facial emotional changes and prefrontal cortex (PFC) activation patterns. Expert evaluations were conducted to assess the outcomes of creative thinking. The results indicated that generative AI significantly enhanced creative thinking performance. Facial emotion analysis revealed that, with generative AI assistance, the brainstorming process generated more fear emotions, while the Theory of Inventive Problem Solving (TRIZ) design process produced more happiness emotions. fNIRS data showed that, with generative AI support, the brainstorming process facilitated activation in the right DLPFC, while the TRIZ design process activated both the left and right DLPFC areas. Machine learning classifiers indicated that facial emotion and fNIRS data could serve as effective indicators for assessing creative thinking performance. The CatBoost classifier achieved an accuracy rate of 91.40 %/89.06 % in the two groups. This study focuses on learners’ facial emotions and PFC activity, revealing that while generative AI enhances creative thinking performance, it may also increase negative emotions. The findings call for caution in using generative AI in creativity education to avoid potential negative psychological effects on students, despite its benefits in promoting creative thinking.
{"title":"Generative AI: A double-edged sword for creative thinking learning — Evidence from facial expressions and fNIRS","authors":"Xinheng Song , Yue Zhang , Zhaolin Lu , Linci Xu , Hengheng Shen","doi":"10.1016/j.compedu.2026.105578","DOIUrl":"10.1016/j.compedu.2026.105578","url":null,"abstract":"<div><div>With the widespread integration of generative AI tools into educational contexts, understanding their influence on learners’ cognitive and emotional processes has become increasingly critical. While AI holds potential for enhancing creativity, its double-edged impact on neurocognitive and emotional processes still requires further investigation. This study investigates the impact of generative AI-based learning tools on the creative thinking learning process. Participants were divided into two groups: a generative AI design group and a traditional design group. They completed tasks employing the divergent brainstorming creative method and the structured innovation TRIZ method. During these tasks, both facial expressions and functional near-infrared spectroscopy (fNIRS) data were collected to explore the effects of generative AI-assisted creative thinking education on students’ facial emotional changes and prefrontal cortex (PFC) activation patterns. Expert evaluations were conducted to assess the outcomes of creative thinking. The results indicated that generative AI significantly enhanced creative thinking performance. Facial emotion analysis revealed that, with generative AI assistance, the brainstorming process generated more fear emotions, while the Theory of Inventive Problem Solving (TRIZ) design process produced more happiness emotions. fNIRS data showed that, with generative AI support, the brainstorming process facilitated activation in the right DLPFC, while the TRIZ design process activated both the left and right DLPFC areas. Machine learning classifiers indicated that facial emotion and fNIRS data could serve as effective indicators for assessing creative thinking performance. The CatBoost classifier achieved an accuracy rate of 91.40 %/89.06 % in the two groups. This study focuses on learners’ facial emotions and PFC activity, revealing that while generative AI enhances creative thinking performance, it may also increase negative emotions. The findings call for caution in using generative AI in creativity education to avoid potential negative psychological effects on students, despite its benefits in promoting creative thinking.</div></div>","PeriodicalId":10568,"journal":{"name":"Computers & Education","volume":"247 ","pages":"Article 105578"},"PeriodicalIF":10.5,"publicationDate":"2026-01-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146001485","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}
Pub Date : 2026-01-18DOI: 10.1016/j.compedu.2026.105576
Chenyu Hou , Gaoxia Zhu , Yanzhi Liu , Vidya Sudarshan , Josephine Leng Leng Chong , Fannie Yifan Zhang , Michael Yong Heng Tan , Yew Soon Ong
As Generative AI becomes increasingly used in various educational contexts, understanding how students engage with these tools during collaborative problem-solving is critical. While prior research suggests that critical thinking is essential in human-AI problem-solving, few studies have examined how instructional interventions, targeting critical thinking, might shape their reliance behaviors and collaborative outcomes. This study investigates the effects of a critical thinking intervention embedded in a problem-based learning (PBL) environment where students are engaged with Generative AI. The intervention combined strategies that foster critical thinking, including authentic instruction, structured dialogue, and AI-supported peer mentoring, aiming to promote students' thoughtful engagement and improve problem-solving performance. Participants (N = 226) were assigned to experimental (with critical thinking interventions) or comparison (without critical thinking interventions) conditions. We used pre- and post-surveys to measure participants' trust, critical thinking, and AI reliance behaviors, and group reports and chat histories to assess their problem-solving quality and creativity. Results revealed that the intervention did not produce significant improvement in self-reported critical thinking, possibly due to the short intervention duration. However, the intervention led to a marginal reduction in students' thoughtless use of Generative AI and significantly reduced the direct adoption of AI-generated content. Notably, students in the intervention condition produced more creative solutions, demonstrating higher levels of originality and idea density in their group reports. These findings suggest that how students use Generative AI is critical, especially when it is almost impossible to control whether they use it or not. The study highlights the importance of designing interventions that cultivate students’ critical thinking to support creative human-AI problem-solving.
{"title":"The effects of critical thinking intervention on reliance behaviors, problem-solving quality, and creativity during human-Generative AI collaborative learning","authors":"Chenyu Hou , Gaoxia Zhu , Yanzhi Liu , Vidya Sudarshan , Josephine Leng Leng Chong , Fannie Yifan Zhang , Michael Yong Heng Tan , Yew Soon Ong","doi":"10.1016/j.compedu.2026.105576","DOIUrl":"10.1016/j.compedu.2026.105576","url":null,"abstract":"<div><div>As Generative AI becomes increasingly used in various educational contexts, understanding how students engage with these tools during collaborative problem-solving is critical. While prior research suggests that critical thinking is essential in human-AI problem-solving, few studies have examined how instructional interventions, targeting critical thinking, might shape their reliance behaviors and collaborative outcomes. This study investigates the effects of a critical thinking intervention embedded in a problem-based learning (PBL) environment where students are engaged with Generative AI. The intervention combined strategies that foster critical thinking, including authentic instruction, structured dialogue, and AI-supported peer mentoring, aiming to promote students' thoughtful engagement and improve problem-solving performance. Participants (N = 226) were assigned to experimental (with critical thinking interventions) or comparison (without critical thinking interventions) conditions. We used pre- and post-surveys to measure participants' trust, critical thinking, and AI reliance behaviors, and group reports and chat histories to assess their problem-solving quality and creativity. Results revealed that the intervention did not produce significant improvement in self-reported critical thinking, possibly due to the short intervention duration. However, the intervention led to a marginal reduction in students' thoughtless use of Generative AI and significantly reduced the direct adoption of AI-generated content. Notably, students in the intervention condition produced more creative solutions, demonstrating higher levels of originality and idea density in their group reports. These findings suggest that <em>how</em> students use Generative AI is critical, especially when it is almost impossible to control <em>whether</em> they use it or not. The study highlights the importance of designing interventions that cultivate students’ critical thinking to support creative human-AI problem-solving.</div></div>","PeriodicalId":10568,"journal":{"name":"Computers & Education","volume":"247 ","pages":"Article 105576"},"PeriodicalIF":10.5,"publicationDate":"2026-01-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145995472","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}
Pub Date : 2026-01-14DOI: 10.1016/j.compedu.2025.105553
Jonas De Bruyne , Charlotte Larmuseau , Lieven De Marez , Durk Talsma , Klaas Bombeke
‘Learning by doing’, or experiential learning, is increasingly implemented through immersive media such as virtual reality (VR) across domains like education and professional training. Immersive technologies enable dynamic instruction and guidance, but this potential remains underexplored. To support learning, cognitive load theory promotes signaling to reduce cognitive load by guiding attention to essential content, while discovery learning encourages minimal guidance to foster exploration. While the temporal aspect of the signaling principle is underrepresented in literature, this study suggests that striking a balance between the theoretical approaches is possible by delaying additional guidance. This work therefore investigates the impact of delayed signaling on experiential learning in VR, using a VR training module on electrofusion welding that is currently used in industry. When comparing performance after training either with immediate or delayed signaling, the data suggested improved procedural learning when signaling was delayed, with an average improvement of 8% in task completion time ( .05, d = .76). Furthermore, the method with delayed signaling did not increase cognitive load, as measured by self-reports, suggesting that discovery learning in combination with (delayed) guidance does not place undue cognitive demand on participants. The findings stress the – currently underexposed – importance of timing of visual aids through signaling and how they can be used to optimize training effectiveness. The results are interpreted in light of existing learning literature with future directions for adaptive training systems highlighted.
“边做边学”或体验式学习越来越多地通过虚拟现实（VR）等沉浸式媒体在教育和专业培训等领域实现。沉浸式技术可以实现动态教学和指导，但这种潜力仍未得到充分开发。为了支持学习，认知负荷理论提倡通过引导注意力到基本内容来减少认知负荷，而发现学习则鼓励最少的指导来促进探索。虽然信号传导原理的时间方面在文献中代表性不足，但本研究表明，通过延迟额外的指导，可以在理论方法之间取得平衡。因此，本研究利用目前在工业中使用的电熔焊接VR培训模块，研究延迟信号对VR体验学习的影响。当比较即时或延迟信号训练后的表现时，数据表明延迟信号时程序学习得到改善，任务完成时间平均提高8% （p < 0.05, d = 0.76）。此外，根据自我报告的测量，延迟信号的方法并没有增加认知负荷，这表明发现学习与（延迟）指导相结合不会对参与者产生不适当的认知需求。研究结果强调了——目前尚未充分暴露的——通过信号提供视觉辅助的时机的重要性，以及如何利用它们来优化训练效果。根据现有的学习文献对结果进行了解释，并强调了自适应训练系统的未来方向。
{"title":"Timing matters! Using delayed signaling to improve experiential learning in procedural VR training","authors":"Jonas De Bruyne , Charlotte Larmuseau , Lieven De Marez , Durk Talsma , Klaas Bombeke","doi":"10.1016/j.compedu.2025.105553","DOIUrl":"10.1016/j.compedu.2025.105553","url":null,"abstract":"<div><div>‘Learning by doing’, or experiential learning, is increasingly implemented through immersive media such as virtual reality (VR) across domains like education and professional training. Immersive technologies enable dynamic instruction and guidance, but this potential remains underexplored. To support learning, cognitive load theory promotes signaling to reduce cognitive load by guiding attention to essential content, while discovery learning encourages minimal guidance to foster exploration. While the temporal aspect of the signaling principle is underrepresented in literature, this study suggests that striking a balance between the theoretical approaches is possible by delaying additional guidance. This work therefore investigates the impact of delayed signaling on experiential learning in VR, using a VR training module on electrofusion welding that is currently used in industry. When comparing performance after training either with immediate or delayed signaling, the data suggested improved procedural learning when signaling was delayed, with an average improvement of 8% in task completion time (<span><math><mi>p</mi></math></span> <span><math><mo><</mo></math></span> .05, <em>d</em> = .76). Furthermore, the method with delayed signaling did not increase cognitive load, as measured by self-reports, suggesting that discovery learning in combination with (delayed) guidance does not place undue cognitive demand on participants. The findings stress the – currently underexposed – importance of timing of visual aids through signaling and how they can be used to optimize training effectiveness. The results are interpreted in light of existing learning literature with future directions for adaptive training systems highlighted.</div></div>","PeriodicalId":10568,"journal":{"name":"Computers & Education","volume":"246 ","pages":"Article 105553"},"PeriodicalIF":10.5,"publicationDate":"2026-01-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145975569","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}
Pub Date : 2026-01-14DOI: 10.1016/j.compedu.2026.105575
Joseph G. Guerriero , Pejman Sajjadi , Janet K. Swim , Alexander Klippel , Jamie DeCoster , Mahda M. Bagher
Virtual reality (VR) serious games can expose people to environmental processes they would not otherwise experience. This can make topics in environmental science more concrete to learners, improving learning outcomes and downstream behaviors related to environmental sustainability. In a randomized experiment (N = 189), we examined the effectiveness of a VR serious game designed to teach people about a topic in environmental science—the Critical Zone—by comparing it to a non-VR version of the game and to a static presentation of the same information on a website. Although the VR serious game promoted greater spatial presence and feelings of awe (which, in turn, translated to feeling more connected with nature), these effects did not translate to improved learning outcomes and pro-environmental policy support as we hypothesized across two separate models. Yet, exploratory analyses revealed a very small but significant indirect pathway by which the VR serious game promoted systems thinking about the Food-Energy-Water (FEW) nexus and pro-environmental policy support: VR (vs other learning formats) led to increases in a sense of spatial presence, then to perceived learning effectiveness, then to FEW systems thinking, and, finally, to pro-environmental policy support. Our results shed light on the mixed effect of VR and spatial presence on learning outcomes discussed in the wider literature on VR in education. Although the original hypotheses were largely unsupported, by exploring and highlighting pathways from learning formats to outcomes, we demonstrate the potential of VR for promoting learning and pro-environmental policy support.
{"title":"Virtual reality serious games for promoting environmental systems thinking and pro-environmental policy support","authors":"Joseph G. Guerriero , Pejman Sajjadi , Janet K. Swim , Alexander Klippel , Jamie DeCoster , Mahda M. Bagher","doi":"10.1016/j.compedu.2026.105575","DOIUrl":"10.1016/j.compedu.2026.105575","url":null,"abstract":"<div><div>Virtual reality (VR) serious games can expose people to environmental processes they would not otherwise experience. This can make topics in environmental science more concrete to learners, improving learning outcomes and downstream behaviors related to environmental sustainability. In a randomized experiment (<em>N</em> = 189), we examined the effectiveness of a VR serious game designed to teach people about a topic in environmental science—the Critical Zone—by comparing it to a non-VR version of the game and to a static presentation of the same information on a website. Although the VR serious game promoted greater spatial presence and feelings of awe (which, in turn, translated to feeling more connected with nature), these effects did not translate to improved learning outcomes and pro-environmental policy support as we hypothesized across two separate models. Yet, exploratory analyses revealed a very small but significant indirect pathway by which the VR serious game promoted systems thinking about the Food-Energy-Water (FEW) nexus and pro-environmental policy support: VR (vs other learning formats) led to increases in a sense of spatial presence, then to perceived learning effectiveness, then to FEW systems thinking, and, finally, to pro-environmental policy support. Our results shed light on the mixed effect of VR and spatial presence on learning outcomes discussed in the wider literature on VR in education. Although the original hypotheses were largely unsupported, by exploring and highlighting pathways from learning formats to outcomes, we demonstrate the potential of VR for promoting learning and pro-environmental policy support.</div></div>","PeriodicalId":10568,"journal":{"name":"Computers & Education","volume":"247 ","pages":"Article 105575"},"PeriodicalIF":10.5,"publicationDate":"2026-01-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145962598","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}
Pub Date : 2026-01-13DOI: 10.1016/j.compedu.2026.105572
Yimin Ning , Wee Tiong Seah , Jihe Chen , Jinhai Liu , Peifen Tan
Feedback is crucial for teacher professional development, yet most studies focus on a single source (e.g., experts or peers). Longitudinal comparisons across sources and analyses of underlying mechanisms are rare. This study therefore investigates how different feedback sources influence teacher learning outcomes. A total of 522 teachers participated in a 13-week professional learning program, including three weeks of reflective practice. Participants were divided into three groups: NF (no external feedback), EF (expert feedback), and AF (AI feedback). Teaching videos, instructional design artifacts, and reflective journals were analyzed to assess outcomes. Teachers in the EF group achieved significantly higher learning outcomes than those in the AF group, who, in turn, outperformed the NF group. These findings underscore the distinctive value of expert feedback in providing depth of insight and contextual sensitivity, while also indicating that AI feedback, although timely and adaptive, cannot fully replace expert judgment. Cross-Lagged Panel Network (CLPN) analysis identified key behaviors and sequences across groups and revealed detailed temporal patterns within behavioral clusters. Qualitative interviews further demonstrated that feedback effectiveness is shaped by the interaction of three dimensions—Time, Object, and Level (TOL)—which form the basis of a framework with 16 elements. The study highlights the complementary strengths of AI adaptability and expert insight, suggesting that multi-source feedback enhances teacher professional development by integrating process support with higher-order regulation.
{"title":"A comparative study of expert, AI, and no external feedback on mathematics teacher learning outcomes in reflective practice","authors":"Yimin Ning , Wee Tiong Seah , Jihe Chen , Jinhai Liu , Peifen Tan","doi":"10.1016/j.compedu.2026.105572","DOIUrl":"10.1016/j.compedu.2026.105572","url":null,"abstract":"<div><div>Feedback is crucial for teacher professional development, yet most studies focus on a single source (e.g., experts or peers). Longitudinal comparisons across sources and analyses of underlying mechanisms are rare. This study therefore investigates how different feedback sources influence teacher learning outcomes. A total of 522 teachers participated in a 13-week professional learning program, including three weeks of reflective practice. Participants were divided into three groups: NF (no external feedback), EF (expert feedback), and AF (AI feedback). Teaching videos, instructional design artifacts, and reflective journals were analyzed to assess outcomes. Teachers in the EF group achieved significantly higher learning outcomes than those in the AF group, who, in turn, outperformed the NF group. These findings underscore the distinctive value of expert feedback in providing depth of insight and contextual sensitivity, while also indicating that AI feedback, although timely and adaptive, cannot fully replace expert judgment. Cross-Lagged Panel Network (CLPN) analysis identified key behaviors and sequences across groups and revealed detailed temporal patterns within behavioral clusters. Qualitative interviews further demonstrated that feedback effectiveness is shaped by the interaction of three dimensions—Time, Object, and Level (TOL)—which form the basis of a framework with 16 elements. The study highlights the complementary strengths of AI adaptability and expert insight, suggesting that multi-source feedback enhances teacher professional development by integrating process support with higher-order regulation.</div></div>","PeriodicalId":10568,"journal":{"name":"Computers & Education","volume":"246 ","pages":"Article 105572"},"PeriodicalIF":10.5,"publicationDate":"2026-01-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145962558","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}
Pub Date : 2026-01-13DOI: 10.1016/j.compedu.2026.105574
Fan Ouyang , Xianping Bai
Multimodal learning analytics (MMLA) has provided new perspectives for computer-supported collaborative learning (CSCL) by capturing multimodal data to explore behavior, social interaction, cognition, regulation, and emotion in CSCL process. However, there are critical challenges in handling multimodal data in CSCL context, such as multimodal data preprocessing methods, selecting suitable analysis methods and tools, and integrating multi-source, multimodal data to represent learning indicators in CSCL process. To fill these gaps, this systematic review constructed a conceptual framework of MMLA in CSCL and provided an overview of the contexts, multimodal data, indicators, data preprocessing methods, analysis methods, and tools, and effects of MMLA applications in CSCL from 2012 to 2024. One hundred fourteen studies articles were included for the final synthesis. Results found that: (1) existing studies primarily focused on groups’ social interactions in CSCL; (2) visual data was commonly adopted in CSCL; (3) the relationships between multimodal data and learning indicators in CSCL included four types, namely One-to-One, Many-to-One, One-to-Many, and Many-to-Many, with particular emphasis on Many-to-One relationships; (4) the most frequently used data preprocessing method was manual coding and extraction, and the utilization of traditional analysis methods (e.g., statistical analysis) had gradually decreased in CSCL, while advanced analysis techniques (e.g., AI algorithms) were gradually gaining traction but were not yet widely adopted; and (5) the application of MMLA in CSCL had positive effects on both learners and instructors, which primarily help instructors comprehensively understanding the CSCL process. Based on the results, this research proposed theoretical, technological, and practical implications to guide future research in the application of MMLA within CSCL contexts.
{"title":"A systematic review of multimodal learning analytics in computer-supported collaborative learning","authors":"Fan Ouyang , Xianping Bai","doi":"10.1016/j.compedu.2026.105574","DOIUrl":"10.1016/j.compedu.2026.105574","url":null,"abstract":"<div><div>Multimodal learning analytics (MMLA) has provided new perspectives for computer-supported collaborative learning (CSCL) by capturing multimodal data to explore behavior, social interaction, cognition, regulation, and emotion in CSCL process. However, there are critical challenges in handling multimodal data in CSCL context, such as multimodal data preprocessing methods, selecting suitable analysis methods and tools, and integrating multi-source, multimodal data to represent learning indicators in CSCL process. To fill these gaps, this systematic review constructed a conceptual framework of MMLA in CSCL and provided an overview of the contexts, multimodal data, indicators, data preprocessing methods, analysis methods, and tools, and effects of MMLA applications in CSCL from 2012 to 2024. One hundred fourteen studies articles were included for the final synthesis. Results found that: (1) existing studies primarily focused on groups’ social interactions in CSCL; (2) visual data was commonly adopted in CSCL; (3) the relationships between multimodal data and learning indicators in CSCL included four types, namely One-to-One, Many-to-One, One-to-Many, and Many-to-Many, with particular emphasis on Many-to-One relationships; (4) the most frequently used data preprocessing method was manual coding and extraction, and the utilization of traditional analysis methods (e.g., statistical analysis) had gradually decreased in CSCL, while advanced analysis techniques (e.g., AI algorithms) were gradually gaining traction but were not yet widely adopted; and (5) the application of MMLA in CSCL had positive effects on both learners and instructors, which primarily help instructors comprehensively understanding the CSCL process. Based on the results, this research proposed theoretical, technological, and practical implications to guide future research in the application of MMLA within CSCL contexts.</div></div>","PeriodicalId":10568,"journal":{"name":"Computers & Education","volume":"247 ","pages":"Article 105574"},"PeriodicalIF":10.5,"publicationDate":"2026-01-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145961761","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}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：