Journal of Computer Assisted Learning最新文献_第4页

Investigating the Influence of Training Difficulty on the Learning Outcomes of Medical Students 训练难度对医学生学习效果的影响

IF 4.6 2区教育学 Q1 EDUCATION & EDUCATIONAL RESEARCH

Journal of Computer Assisted Learning

Pub Date : 2025-12-11 DOI: 10.1002/jcal.70172

Erva Nihan Kandemir, Jill-Jênn Vie, Adam Sanchez-Ayte, Olivier Palombi, Franck Ramus

Background

Determining an optimal training difficulty level for the best learning outcome is a crucial goal for adaptive educational systems. The literature supports the Inverted U-shape Hypothesis, suggesting that the ideal challenge level for learning is neither too easy nor too difficult. However, this optimal point depends on the type of training and response modality and may vary across domains, necessitating thorough examination before implementing adaptive learning procedures.

Objectives

This study aimed to investigate the influence of training difficulty on the learning outcomes of French medical students.

Methods

Using data from a national educational platform, we explored the influence of the mean question difficulty encountered during training, relative to individual student ability, on the learning outcomes of medical students across diverse medical specialties. Importantly, the mean difficulty level varied randomly between students on this platform, mirroring a quasi-experimental design and enabling a thorough exploration of these effects. We first employed the Elo rating system to estimate the difficulty of platform questions and the evolution of students' abilities. A linear mixed-effects model was then used, with final exam performance as the main outcome and mean relative question difficulty during training (linear and quadratic terms) as the main predictor.

Results and Conclusions

Results showed a significant negative quadratic effect of mean relative difficulty on the final exam performance, revealing optimal difficulty levels for each medical specialty. Additionally, the analysis demonstrated that students with high abilities displayed a more pronounced inverted U-shaped relationship between training difficulty and final exam scores. This study advances our understanding of optimal training difficulty in the complex realm of medical education, by emphasising the need to acknowledge variability across medical specialties and student abilities.

为获得最佳学习效果而确定最佳训练难度水平是适应性教育系统的一个重要目标。文献支持倒u形假说，表明学习的理想挑战水平既不太容易也不太困难。然而，这个最佳点取决于训练类型和反应方式，并且可能在不同的领域有所不同，因此在实施适应性学习程序之前需要进行彻底的检查。目的探讨训练难度对法国医学生学习效果的影响。方法利用国家教育平台的数据，探讨不同医学专业医学生在培训过程中遇到的平均问题难度相对于学生个人能力对学习结果的影响。重要的是，该平台上学生的平均难度水平随机变化，反映了准实验设计，可以深入探索这些影响。我们首先使用Elo评分系统来评估平台问题的难度和学生能力的演变。然后使用线性混合效应模型，以期末考试成绩为主要结果，训练期间的平均相对问题难度（线性和二次项）为主要预测因子。结果与结论平均相对难度对期末考试成绩有显著的负二次效应，揭示了各医学专业的最佳难度水平。此外，分析表明，高能力学生在训练难度和期末考试成绩之间表现出更明显的倒u型关系。本研究通过强调承认医学专业和学生能力差异的必要性，提高了我们对复杂医学教育领域中最佳训练难度的理解。

{"title":"Investigating the Influence of Training Difficulty on the Learning Outcomes of Medical Students","authors":"Erva Nihan Kandemir, Jill-Jênn Vie, Adam Sanchez-Ayte, Olivier Palombi, Franck Ramus","doi":"10.1002/jcal.70172","DOIUrl":"https://doi.org/10.1002/jcal.70172","url":null,"abstract":"<div>\u0000 \u0000 \u0000 <section>\u0000 \u0000 <h3> Background</h3>\u0000 \u0000 <p>Determining an optimal training difficulty level for the best learning outcome is a crucial goal for adaptive educational systems. The literature supports the Inverted <i>U</i>-shape Hypothesis, suggesting that the ideal challenge level for learning is neither too easy nor too difficult. However, this optimal point depends on the type of training and response modality and may vary across domains, necessitating thorough examination before implementing adaptive learning procedures.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Objectives</h3>\u0000 \u0000 <p>This study aimed to investigate the influence of training difficulty on the learning outcomes of French medical students.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Methods</h3>\u0000 \u0000 <p>Using data from a national educational platform, we explored the influence of the mean question difficulty encountered during training, relative to individual student ability, on the learning outcomes of medical students across diverse medical specialties. Importantly, the mean difficulty level varied randomly between students on this platform, mirroring a quasi-experimental design and enabling a thorough exploration of these effects. We first employed the Elo rating system to estimate the difficulty of platform questions and the evolution of students' abilities. A linear mixed-effects model was then used, with final exam performance as the main outcome and mean relative question difficulty during training (linear and quadratic terms) as the main predictor.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Results and Conclusions</h3>\u0000 \u0000 <p>Results showed a significant negative quadratic effect of mean relative difficulty on the final exam performance, revealing optimal difficulty levels for each medical specialty. Additionally, the analysis demonstrated that students with high abilities displayed a more pronounced inverted <i>U</i>-shaped relationship between training difficulty and final exam scores. This study advances our understanding of optimal training difficulty in the complex realm of medical education, by emphasising the need to acknowledge variability across medical specialties and student abilities.</p>\u0000 </section>\u0000 </div>","PeriodicalId":48071,"journal":{"name":"Journal of Computer Assisted Learning","volume":"42 1","pages":""},"PeriodicalIF":4.6,"publicationDate":"2025-12-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145739516","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"教育学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Effects of a Self-Regulated Cooperative Programming Approach on Elementary Students' Programming Knowledge, Metacognitive Awareness, and Self-Efficacy 自我调节合作编程方式对小学生编程知识、元认知意识和自我效能感的影响

IF 4.6 2区教育学 Q1 EDUCATION & EDUCATIONAL RESEARCH

Journal of Computer Assisted Learning

Pub Date : 2025-12-09 DOI: 10.1002/jcal.70167

Chengye Liu, Zhihao Cui, Xiaojing Weng

Background

Programming is a vital digital skill that can equip students with the necessary competencies for future employment opportunities. However, due to inadequate metacognitive awareness, self-regulation abilities, and a lack of prior programming knowledge, elementary school students may struggle with completing complex programming tasks.

Objectives

We aim to propose a self-regulated cooperative programming (SR-CoP) approach. While exploring the effectiveness of the approach in students' programming knowledge, metacognitive awareness, and self-efficacy.

Methods

Participants were 113 fourth graders (54 males, 59 females). Their ages ranged from 10 to 11 years. They were randomly assigned to three groups: the control group using individual programming (C-IdP), experimental group 1 employing SR-CoP, and experimental group 2 employing conventional cooperative programming (C-CoP).

Results

The SR-CoP approach significantly enhanced students' metacognitive awareness and self-efficacy compared to the other groups. Regarding programming knowledge, no statistically significant difference were found among the three groups, suggesting that SR-CoP supports non-cognitive outcomes without compromising domain knowledge acquisition.

Conclusions

By employing self-regulation strategies, students engage in planning, self-monitoring, and self-reflection during project completion and cooperative programming activities. This practice is advantageous for enhancing elementary students' metacognitive awareness and self-efficacy.

编程是一项重要的数字技能，可以使学生具备未来就业机会所需的能力。然而，由于元认知意识和自我调节能力的不足，以及缺乏事先的编程知识，小学生在完成复杂的编程任务时可能会遇到困难。目的提出一种自我调节的协同规划（SR-CoP）方法。同时探讨该方法对学生编程知识、元认知意识和自我效能感的影响。方法调查对象为113名四年级学生（男54名，女59名）。他们的年龄从10岁到11岁不等。随机分为3组：对照组采用个体规划（C-IdP），实验组1采用SR-CoP，实验组2采用常规合作规划（C-CoP）。结果SR-CoP方法显著提高了学生的元认知意识和自我效能感。在编程知识方面，三组之间没有统计学差异，这表明SR-CoP支持非认知结果，而不影响领域知识的获取。通过采用自我调节策略，学生在项目完成和合作编程活动中进行计划、自我监控和自我反思。这有利于提高小学生的元认知意识和自我效能感。

{"title":"Effects of a Self-Regulated Cooperative Programming Approach on Elementary Students' Programming Knowledge, Metacognitive Awareness, and Self-Efficacy","authors":"Chengye Liu, Zhihao Cui, Xiaojing Weng","doi":"10.1002/jcal.70167","DOIUrl":"https://doi.org/10.1002/jcal.70167","url":null,"abstract":"<div>\u0000 \u0000 \u0000 <section>\u0000 \u0000 <h3> Background</h3>\u0000 \u0000 <p>Programming is a vital digital skill that can equip students with the necessary competencies for future employment opportunities. However, due to inadequate metacognitive awareness, self-regulation abilities, and a lack of prior programming knowledge, elementary school students may struggle with completing complex programming tasks.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Objectives</h3>\u0000 \u0000 <p>We aim to propose a self-regulated cooperative programming (SR-CoP) approach. While exploring the effectiveness of the approach in students' programming knowledge, metacognitive awareness, and self-efficacy.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Methods</h3>\u0000 \u0000 <p>Participants were 113 fourth graders (54 males, 59 females). Their ages ranged from 10 to 11 years. They were randomly assigned to three groups: the control group using individual programming (C-IdP), experimental group 1 employing SR-CoP, and experimental group 2 employing conventional cooperative programming (C-CoP).</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Results</h3>\u0000 \u0000 <p>The SR-CoP approach significantly enhanced students' metacognitive awareness and self-efficacy compared to the other groups. Regarding programming knowledge, no statistically significant difference were found among the three groups, suggesting that SR-CoP supports non-cognitive outcomes without compromising domain knowledge acquisition.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Conclusions</h3>\u0000 \u0000 <p>By employing self-regulation strategies, students engage in planning, self-monitoring, and self-reflection during project completion and cooperative programming activities. This practice is advantageous for enhancing elementary students' metacognitive awareness and self-efficacy.</p>\u0000 </section>\u0000 </div>","PeriodicalId":48071,"journal":{"name":"Journal of Computer Assisted Learning","volume":"42 1","pages":""},"PeriodicalIF":4.6,"publicationDate":"2025-12-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://onlinelibrary.wiley.com/doi/epdf/10.1002/jcal.70167","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145750875","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"教育学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Systematically Reviewing the Rigour of Immersive Virtual Reality Research in STEM Education: A Deep Dive Into Threats to Internal Validity 系统回顾STEM教育中沉浸式虚拟现实研究的严谨性：深入探讨对内部有效性的威胁

IF 4.6 2区教育学 Q1 EDUCATION & EDUCATIONAL RESEARCH

Journal of Computer Assisted Learning

Pub Date : 2025-12-09 DOI: 10.1002/jcal.70165

Amédee Marchand Martella, Alyssa P. Lawson, Ronald C. Martella, Kristen LaBonte, Fangzheng Zhao, Cynthia Y. Delgado, Mitchell E. Munns, Ashleigh Wells LeRoy, Justin A. Gluck, Richard E. Mayer

Background

There have been many initiatives focused on implementing IVR in the classroom to either replace or supplement conventional instruction. The efficacy of these initiatives is often informed by IVR media comparison studies that examine the learning outcomes of students who learn academic content in IVR versus conventional media. However, the current methodological rigour of research on the use of IVR for learning in STEM education has not yet been established to ensure recommendations concerning its use in education are rooted in rigorous research studies.

Objective

The present review aimed to fill this gap in the literature by evaluating published journal articles, conference proceedings, and dissertations (between the years 2013 and 2023) related to IVR comparison studies in STEM education.

Method

The 44 articles were evaluated with respect to 13 internal validity controls (i.e., research design, justified sample size, group equivalence, attrition equivalence, operational definitions provided, matched dosage, matched timeframe, matched content, implementation fidelity, practice equivalence with dependent measures [DMs], equivalent DMs, interrater agreement for DM scoring and technical adequacy of DMs).

Results

Results indicated there were no articles that explicitly met all 13 internal validity controls; the average number of control issues was 5.36. When articles were not penalised for missing information, 21 articles adhered to all internal validity controls. As can be seen when comparing these findings, the lack of methodological information was a glaring problem.

Conclusion

These findings indicate that there are both methodological problems and reporting problems in IVR media comparison research that need to be addressed to advance future IVR research. Recommendations for productively moving forward are discussed.

已经有许多倡议集中在课堂上实施IVR，以取代或补充传统教学。这些举措的有效性通常通过IVR媒体比较研究来了解，这些研究检查了在IVR和传统媒体中学习学术内容的学生的学习成果。然而，目前关于在STEM教育中使用IVR进行学习的严谨研究方法尚未建立，以确保有关其在教育中使用的建议植根于严谨的研究。本综述旨在通过评估已发表的与STEM教育中IVR比较研究相关的期刊文章、会议论文集和论文（2013年至2023年）来填补这一文献空白。方法对44篇文献进行13项内部效度控制(即研究设计、合理样本量、分组等效、消耗等效、提供的操作定义、匹配剂量、匹配时间框架、匹配内容、实施保真度、实践与依赖测量的等效性、等效DMs、口译员对DM评分和DM技术充分性的协议)。结果没有文章明确满足全部13项内部效度控制；平均控制问题数为5.36个。当文章没有因遗漏信息而受到处罚时，有21篇文章遵守了所有内部有效性控制。在比较这些调查结果时可以看出，缺乏方法学信息是一个明显的问题。这些发现表明，在IVR媒体比较研究中存在方法问题和报告问题，需要解决这些问题，以推进未来的IVR研究。讨论了有效推进的建议。

{"title":"Systematically Reviewing the Rigour of Immersive Virtual Reality Research in STEM Education: A Deep Dive Into Threats to Internal Validity","authors":"Amédee Marchand Martella, Alyssa P. Lawson, Ronald C. Martella, Kristen LaBonte, Fangzheng Zhao, Cynthia Y. Delgado, Mitchell E. Munns, Ashleigh Wells LeRoy, Justin A. Gluck, Richard E. Mayer","doi":"10.1002/jcal.70165","DOIUrl":"https://doi.org/10.1002/jcal.70165","url":null,"abstract":"<div>\u0000 \u0000 \u0000 <section>\u0000 \u0000 <h3> Background</h3>\u0000 \u0000 <p>There have been many initiatives focused on implementing IVR in the classroom to either replace or supplement conventional instruction. The efficacy of these initiatives is often informed by IVR media comparison studies that examine the learning outcomes of students who learn academic content in IVR versus conventional media. However, the current methodological rigour of research on the use of IVR for learning in STEM education has not yet been established to ensure recommendations concerning its use in education are rooted in rigorous research studies.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Objective</h3>\u0000 \u0000 <p>The present review aimed to fill this gap in the literature by evaluating published journal articles, conference proceedings, and dissertations (between the years 2013 and 2023) related to IVR comparison studies in STEM education.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Method</h3>\u0000 \u0000 <p>The 44 articles were evaluated with respect to 13 internal validity controls (i.e., research design, justified sample size, group equivalence, attrition equivalence, operational definitions provided, matched dosage, matched timeframe, matched content, implementation fidelity, practice equivalence with dependent measures [DMs], equivalent DMs, interrater agreement for DM scoring and technical adequacy of DMs).</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Results</h3>\u0000 \u0000 <p>Results indicated there were no articles that explicitly met all 13 internal validity controls; the average number of control issues was 5.36. When articles were not penalised for missing information, 21 articles adhered to all internal validity controls. As can be seen when comparing these findings, the lack of methodological information was a glaring problem.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Conclusion</h3>\u0000 \u0000 <p>These findings indicate that there are both methodological problems and reporting problems in IVR media comparison research that need to be addressed to advance future IVR research. Recommendations for productively moving forward are discussed.</p>\u0000 </section>\u0000 </div>","PeriodicalId":48071,"journal":{"name":"Journal of Computer Assisted Learning","volume":"42 1","pages":""},"PeriodicalIF":4.6,"publicationDate":"2025-12-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://onlinelibrary.wiley.com/doi/epdf/10.1002/jcal.70165","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145750876","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"教育学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Unveiling the Relationship Between Undergraduate Students' Emergent Roles and Learning Performance in Collaborative Argumentation-Based Learning: Insights From Sequence Clustering and Entropy Analysis 揭示大学生协同论证学习中突发角色与学习绩效的关系：来自序列聚类和熵分析的启示

IF 4.6 2区教育学 Q1 EDUCATION & EDUCATIONAL RESEARCH

Journal of Computer Assisted Learning

Pub Date : 2025-12-09 DOI: 10.1002/jcal.70173

Qingtang Liu, Yubei Chang, Xiaojuan Li, Xinxin Zheng

Background

Roles are the socio-cognitive interaction patterns in social learning. Analyzing emergent roles provides insights into how students naturally participate and contribute knowledge in productive collaborative argumentation-based learning (CABLE). However, previous studies lack a comprehensive and in-depth understanding of the emergent roles in CABLE. In addition, the relationship between the emergent roles and the CABLE performance of groups and individuals needs to be further explored.

Objectives

This study aims to identify emergent roles in CABLE and explore the relationship between the dynamic fluctuation of emergent roles and learning performance from both individual and group perspectives.

Methods

To fill this gap, a collaborative argumentation discussion activity was conducted among 33 undergraduate students. The study analyzed the fine-grained collaborative discourse moves based on the community of inquiry (CoI) framework. Then, the study identified emergent roles and quantified the dynamic fluctuation of emergent roles through sequence clustering and entropy analysis. Finally, the relationship between the emergent roles and the CABLE performance of groups and individuals was explored.

Results and Conclusions

The findings revealed that students emerged as Discussion facilitators, Viewpoint contributors, and Knowledge integrators during CABLE. Students with high argumentation skills had significantly higher role entropy values than students with low argumentation skills, and they flexibly regulated and transitioned their emergent roles. Furthermore, the higher the dynamic fluctuation of group emergent roles, the better the group learning performance tended to be. Based on the findings, the study proposed implications for instructional design and future research directions to optimize students' CABLE.

角色是社会学习中的社会认知互动模式。分析紧急角色提供了洞察学生如何自然地参与和贡献知识在富有成效的协作论证为基础的学习（CABLE）。然而，以往的研究对有线电视中的新兴角色缺乏全面而深入的认识。此外，突现角色与群体和个人的CABLE绩效之间的关系还有待进一步探讨。目的本研究旨在从个体和群体两个角度，识别大学生网络学习中的突发角色，并探讨突发角色动态波动与学习绩效的关系。方法针对这一不足，对33名大学生开展了协同论证讨论活动。本研究分析了基于探究社区（CoI）框架的细粒度协作话语移动。在此基础上，通过序列聚类和熵分析，确定了应急角色，量化了应急角色的动态波动。最后，探讨了突发角色与群体和个人的有线电视绩效之间的关系。结果和结论研究结果表明，学生在CABLE期间成为讨论促进者、观点贡献者和知识集成者。论证能力高的学生角色熵值显著高于论证能力低的学生，并且能够灵活调节和转换应急角色。此外，群体涌现角色动态波动越高，群体学习绩效越好。在此基础上，本研究提出了教学设计的启示和未来的研究方向，以优化学生的CABLE。

{"title":"Unveiling the Relationship Between Undergraduate Students' Emergent Roles and Learning Performance in Collaborative Argumentation-Based Learning: Insights From Sequence Clustering and Entropy Analysis","authors":"Qingtang Liu, Yubei Chang, Xiaojuan Li, Xinxin Zheng","doi":"10.1002/jcal.70173","DOIUrl":"https://doi.org/10.1002/jcal.70173","url":null,"abstract":"<div>\u0000 \u0000 \u0000 <section>\u0000 \u0000 <h3> Background</h3>\u0000 \u0000 <p>Roles are the socio-cognitive interaction patterns in social learning. Analyzing emergent roles provides insights into how students naturally participate and contribute knowledge in productive collaborative argumentation-based learning (CABLE). However, previous studies lack a comprehensive and in-depth understanding of the emergent roles in CABLE. In addition, the relationship between the emergent roles and the CABLE performance of groups and individuals needs to be further explored.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Objectives</h3>\u0000 \u0000 <p>This study aims to identify emergent roles in CABLE and explore the relationship between the dynamic fluctuation of emergent roles and learning performance from both individual and group perspectives.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Methods</h3>\u0000 \u0000 <p>To fill this gap, a collaborative argumentation discussion activity was conducted among 33 undergraduate students. The study analyzed the fine-grained collaborative discourse moves based on the community of inquiry (CoI) framework. Then, the study identified emergent roles and quantified the dynamic fluctuation of emergent roles through sequence clustering and entropy analysis. Finally, the relationship between the emergent roles and the CABLE performance of groups and individuals was explored.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Results and Conclusions</h3>\u0000 \u0000 <p>The findings revealed that students emerged as Discussion facilitators, Viewpoint contributors, and Knowledge integrators during CABLE. Students with high argumentation skills had significantly higher role entropy values than students with low argumentation skills, and they flexibly regulated and transitioned their emergent roles. Furthermore, the higher the dynamic fluctuation of group emergent roles, the better the group learning performance tended to be. Based on the findings, the study proposed implications for instructional design and future research directions to optimize students' CABLE.</p>\u0000 </section>\u0000 </div>","PeriodicalId":48071,"journal":{"name":"Journal of Computer Assisted Learning","volume":"42 1","pages":""},"PeriodicalIF":4.6,"publicationDate":"2025-12-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145739488","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"教育学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Are Teachers Addicted to AI? Analysing Factors Influencing Dependence on Generative AI Through the I-PACE Model 教师对人工智能上瘾了吗？通过I-PACE模型分析对生成式人工智能依赖的影响因素

IF 4.6 2区教育学 Q1 EDUCATION & EDUCATIONAL RESEARCH

Journal of Computer Assisted Learning

Pub Date : 2025-12-05 DOI: 10.1002/jcal.70174

Yiran Du, Mi Tang, Kunjie Jia, Chenghao Wang, Bin Zou

Background

The integration of generative artificial intelligence (AI) into education has revolutionised teaching practices, offering educators advanced tools for lesson planning, content creation, personalised learning and administrative automation. While AI enhances efficiency and instructional effectiveness, concerns have emerged regarding teachers' potential overreliance on these technologies, leading to AI addiction.

Objectives

This study applies the I-PACE model (Interaction of Person-Affect-Cognition-Execution) to explore the psychological and behavioural mechanisms underlying teachers' dependence on generative AI.

Methods

Using survey data from 1750 teachers in Huanghua, China, the study examines factors such as self-efficacy, need for cognition, mood regulation, positive affect, perceived usefulness and cognitive absorption in shaping AI addiction.

Results

Findings indicate that cognitive absorption is the strongest predictor of AI dependence, while perceived usefulness, self-efficacy and positive affect contribute indirectly through reinforcement mechanisms. Notably, mood regulation and need for cognition do not significantly influence AI addiction, suggesting that AI engagement in education is driven more by functional efficiency than emotional dependence.

Conclusions

The results highlight the importance of fostering mindful AI integration in teaching to prevent habitual overreliance. This study provides theoretical contributions by extending the I-PACE model to the context of AI addiction in education and offers practical insights for educators, institutions and policymakers in promoting responsible AI use while maintaining teachers' professional autonomy and cognitive engagement.

将生成式人工智能（AI）集成到教育中已经彻底改变了教学实践，为教育工作者提供了课程规划、内容创建、个性化学习和管理自动化的先进工具。虽然人工智能提高了效率和教学效果，但人们担心教师可能过度依赖这些技术，导致人工智能成瘾。目的运用I-PACE模型（人-情感-认知-执行的交互作用）探讨教师对生成式人工智能依赖的心理和行为机制。方法利用黄骅市1750名教师的调查数据，研究了自我效能感、认知需求、情绪调节、积极影响、感知有用性和认知吸收等因素在人工智能成瘾形成中的作用。结果认知吸收是人工智能依赖的最强预测因子，感知有用性、自我效能感和积极情绪通过强化机制间接影响人工智能依赖。值得注意的是，情绪调节和认知需求并没有显著影响人工智能成瘾，这表明人工智能在教育中的参与更多是由功能效率驱动的，而不是情感依赖。结论研究结果强调了在教学中培养正念人工智能整合的重要性，以防止习惯性的过度依赖。本研究通过将I-PACE模型扩展到教育中的人工智能成瘾背景，提供了理论贡献，并为教育工作者、机构和政策制定者提供了实践见解，以促进负责任的人工智能使用，同时保持教师的专业自主权和认知参与。

{"title":"Are Teachers Addicted to AI? Analysing Factors Influencing Dependence on Generative AI Through the I-PACE Model","authors":"Yiran Du, Mi Tang, Kunjie Jia, Chenghao Wang, Bin Zou","doi":"10.1002/jcal.70174","DOIUrl":"https://doi.org/10.1002/jcal.70174","url":null,"abstract":"<div>\u0000 \u0000 \u0000 <section>\u0000 \u0000 <h3> Background</h3>\u0000 \u0000 <p>The integration of generative artificial intelligence (AI) into education has revolutionised teaching practices, offering educators advanced tools for lesson planning, content creation, personalised learning and administrative automation. While AI enhances efficiency and instructional effectiveness, concerns have emerged regarding teachers' potential overreliance on these technologies, leading to AI addiction.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Objectives</h3>\u0000 \u0000 <p>This study applies the I-PACE model (Interaction of Person-Affect-Cognition-Execution) to explore the psychological and behavioural mechanisms underlying teachers' dependence on generative AI.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Methods</h3>\u0000 \u0000 <p>Using survey data from 1750 teachers in Huanghua, China, the study examines factors such as self-efficacy, need for cognition, mood regulation, positive affect, perceived usefulness and cognitive absorption in shaping AI addiction.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Results</h3>\u0000 \u0000 <p>Findings indicate that cognitive absorption is the strongest predictor of AI dependence, while perceived usefulness, self-efficacy and positive affect contribute indirectly through reinforcement mechanisms. Notably, mood regulation and need for cognition do not significantly influence AI addiction, suggesting that AI engagement in education is driven more by functional efficiency than emotional dependence.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Conclusions</h3>\u0000 \u0000 <p>The results highlight the importance of fostering mindful AI integration in teaching to prevent habitual overreliance. This study provides theoretical contributions by extending the I-PACE model to the context of AI addiction in education and offers practical insights for educators, institutions and policymakers in promoting responsible AI use while maintaining teachers' professional autonomy and cognitive engagement.</p>\u0000 </section>\u0000 </div>","PeriodicalId":48071,"journal":{"name":"Journal of Computer Assisted Learning","volume":"42 1","pages":""},"PeriodicalIF":4.6,"publicationDate":"2025-12-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://onlinelibrary.wiley.com/doi/epdf/10.1002/jcal.70174","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145695575","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"教育学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

The Role of Trait Cognitive Flexibility on Game-Based Learning: A Media-Comparison Study 特质认知灵活性在游戏学习中的作用：一项媒介比较研究

IF 4.6 2区教育学 Q1 EDUCATION & EDUCATIONAL RESEARCH

Journal of Computer Assisted Learning

Pub Date : 2025-12-03 DOI: 10.1002/jcal.70169

Fangzheng Zhao, Qiyang Lin, Richard E. Mayer

Background

Previous media comparison research has often compared learning from instructional games with learning from traditional media. However, there is a limited exploration into how the learner's cognitive traits, particularly cognitive flexibility, influence learning outcomes across different media.

Objectives

This study aims to investigate the impact of cognitive flexibility on learning outcomes when using two distinct instructional media: an instructional game and a slideshow lesson.

Methods

A total of 142 undergraduate students participated in this study, engaging in a 1.5-h session focusing on epidemiology. They were randomly assigned to either the instructional game Crystal Island or a slideshow lesson covering the same content. Their trait cognitive flexibility and learning outcomes were evaluated through post-questionnaires and posttests.

Results and Conclusions

The study found that students with higher cognitive flexibility scored better on transfer tests. Additionally, the slides group outperformed the game group. Interestingly, no significant interaction was observed between cognitive flexibility and the assigned learning medium. Overall, this study replicated some findings from previous media comparison studies and highlighted the need to further explore cognitive flexibility as a key individual trait in multimedia learning.

以往的媒介比较研究经常将教学游戏学习与传统媒介学习进行比较。然而，关于学习者的认知特征，特别是认知灵活性如何影响不同媒体的学习结果的探索有限。本研究旨在探讨在使用两种不同的教学媒体：教学游戏和幻灯片课程时，认知灵活性对学习结果的影响。方法对142名大学生进行为期1.5小时的流行病学调查。他们被随机分配到教学游戏《水晶岛》或涵盖相同内容的幻灯片课中。通过问卷调查和后测对学生的认知灵活性和学习效果进行评价。结果与结论研究发现认知灵活性越高的学生在迁移测试中得分越高。此外，幻灯片组的表现优于游戏组。有趣的是，认知灵活性和指定的学习媒介之间没有明显的相互作用。总的来说，本研究重复了先前媒体比较研究的一些发现，并强调了进一步探索认知灵活性作为多媒体学习中关键的个体特征的必要性。

{"title":"The Role of Trait Cognitive Flexibility on Game-Based Learning: A Media-Comparison Study","authors":"Fangzheng Zhao, Qiyang Lin, Richard E. Mayer","doi":"10.1002/jcal.70169","DOIUrl":"https://doi.org/10.1002/jcal.70169","url":null,"abstract":"<div>\u0000 \u0000 \u0000 <section>\u0000 \u0000 <h3> Background</h3>\u0000 \u0000 <p>Previous media comparison research has often compared learning from instructional games with learning from traditional media. However, there is a limited exploration into how the learner's cognitive traits, particularly cognitive flexibility, influence learning outcomes across different media.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Objectives</h3>\u0000 \u0000 <p>This study aims to investigate the impact of cognitive flexibility on learning outcomes when using two distinct instructional media: an instructional game and a slideshow lesson.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Methods</h3>\u0000 \u0000 <p>A total of 142 undergraduate students participated in this study, engaging in a 1.5-h session focusing on epidemiology. They were randomly assigned to either the instructional game <i>Crystal Island</i> or a slideshow lesson covering the same content. Their trait cognitive flexibility and learning outcomes were evaluated through post-questionnaires and posttests.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Results and Conclusions</h3>\u0000 \u0000 <p>The study found that students with higher cognitive flexibility scored better on transfer tests. Additionally, the slides group outperformed the game group. Interestingly, no significant interaction was observed between cognitive flexibility and the assigned learning medium. Overall, this study replicated some findings from previous media comparison studies and highlighted the need to further explore cognitive flexibility as a key individual trait in multimedia learning.</p>\u0000 </section>\u0000 </div>","PeriodicalId":48071,"journal":{"name":"Journal of Computer Assisted Learning","volume":"42 1","pages":""},"PeriodicalIF":4.6,"publicationDate":"2025-12-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145695132","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"教育学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

How LLM Chatbots Shape Self-Regulated Language Learning: The Interplay of Cognitive Load and Basic Psychological Needs LLM聊天机器人如何塑造自我调节的语言学习：认知负荷和基本心理需求的相互作用

IF 4.6 2区教育学 Q1 EDUCATION & EDUCATIONAL RESEARCH

Journal of Computer Assisted Learning

Pub Date : 2025-12-02 DOI: 10.1002/jcal.70163

Chun Hao, Ning Ma, Mohd Rashid Bin Mohd Saad, Huzaina Binti Abdul Halim, Mengmeng Guo, Mengyao Hao

Background

The integration of large language model (LLM) chatbots into language education has opened new opportunities for supporting self-regulated language learning (SRLL). However, their effectiveness depends on how cognitive load and basic psychological needs (BPNs) jointly shape learners' self-regulation.

Objectives

This study proposes an exploratory model to elucidate the potential pathways by which extraneous and germane cognitive load (ECL, GCL) and the satisfaction of BPNs (autonomy, competence, and relatedness) jointly influence SRLL in chatbot-assisted environments.

Methods

A two-stage explanatory mixed-methods design was employed. Quantitative data were collected from 237 English as a foreign language (EFL) learners and analysed using partial least squares structural equation modelling (PLS-SEM). Qualitative data from semi-structured interviews were thematically analysed to explain and triangulate the quantitative findings.

Results and Conclusions

ECL generated by chatbots' technological limitations did not directly undermine SRLL but significantly frustrated perceived competence (PC), which in turn discouraged self-regulation. Nevertheless, these negative effects were limited, as most learners remained willing to engage with LLM chatbots. In contrast, BPN satisfaction emerged as a strong motivator of SRLL behaviours: PC and perceived relatedness (PR) directly promoted SRLL, while perceived autonomy (PA) and PC indirectly facilitated SRLL by enhancing GCL, which also exerted a direct, significant influence on SRLL. These findings highlight the dual role of LLM chatbots: While their limitations may induce tolerable ECL and competence frustration, their affordances in satisfying BPNs and fostering GCL make them a promising tool for sustaining SRLL in EFL contexts.

大型语言模型（LLM）聊天机器人与语言教育的整合为支持自我调节语言学习（SRLL）提供了新的机会。然而，其有效性取决于认知负荷和基本心理需求如何共同塑造学习者的自我调节。本研究提出了一个探索性模型，以阐明在聊天机器人辅助环境中，外部和相关的认知负荷（ECL、GCL）和bpn满意度（自主性、能力和相关性）共同影响SRLL的潜在途径。方法采用两阶段解释混合方法设计。本研究收集了237名英语学习者的定量数据，并使用偏最小二乘结构方程模型（PLS-SEM）进行分析。来自半结构化访谈的定性数据进行了主题分析，以解释和三角测量定量结果。由聊天机器人的技术限制产生的ECL并不会直接破坏SRLL，但会显著阻碍感知能力（PC），进而阻碍自我调节。然而，这些负面影响是有限的，因为大多数学习者仍然愿意参与法学硕士聊天机器人。与此相反，工作伙伴关系满意度是员工自主行为的重要驱动因素，PC和感知相关性（PR）直接促进了员工自主行为，而感知自主性（PA）和PC通过增强员工合作关系间接促进了员工自主行为，二者也对员工自主行为产生了直接且显著的影响。这些发现强调了LLM聊天机器人的双重作用：虽然它们的局限性可能会导致可容忍的ECL和能力挫折，但它们在满足bpn和培养GCL方面的优势使它们成为在英语环境中维持SRLL的有希望的工具。

{"title":"How LLM Chatbots Shape Self-Regulated Language Learning: The Interplay of Cognitive Load and Basic Psychological Needs","authors":"Chun Hao, Ning Ma, Mohd Rashid Bin Mohd Saad, Huzaina Binti Abdul Halim, Mengmeng Guo, Mengyao Hao","doi":"10.1002/jcal.70163","DOIUrl":"https://doi.org/10.1002/jcal.70163","url":null,"abstract":"<div>\u0000 \u0000 \u0000 <section>\u0000 \u0000 <h3> Background</h3>\u0000 \u0000 <p>The integration of large language model (LLM) chatbots into language education has opened new opportunities for supporting self-regulated language learning (SRLL). However, their effectiveness depends on how cognitive load and basic psychological needs (BPNs) jointly shape learners' self-regulation.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Objectives</h3>\u0000 \u0000 <p>This study proposes an exploratory model to elucidate the potential pathways by which extraneous and germane cognitive load (ECL, GCL) and the satisfaction of BPNs (autonomy, competence, and relatedness) jointly influence SRLL in chatbot-assisted environments.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Methods</h3>\u0000 \u0000 <p>A two-stage explanatory mixed-methods design was employed. Quantitative data were collected from 237 English as a foreign language (EFL) learners and analysed using partial least squares structural equation modelling (PLS-SEM). Qualitative data from semi-structured interviews were thematically analysed to explain and triangulate the quantitative findings.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Results and Conclusions</h3>\u0000 \u0000 <p>ECL generated by chatbots' technological limitations did not directly undermine SRLL but significantly frustrated perceived competence (PC), which in turn discouraged self-regulation. Nevertheless, these negative effects were limited, as most learners remained willing to engage with LLM chatbots. In contrast, BPN satisfaction emerged as a strong motivator of SRLL behaviours: PC and perceived relatedness (PR) directly promoted SRLL, while perceived autonomy (PA) and PC indirectly facilitated SRLL by enhancing GCL, which also exerted a direct, significant influence on SRLL. These findings highlight the dual role of LLM chatbots: While their limitations may induce tolerable ECL and competence frustration, their affordances in satisfying BPNs and fostering GCL make them a promising tool for sustaining SRLL in EFL contexts.</p>\u0000 </section>\u0000 </div>","PeriodicalId":48071,"journal":{"name":"Journal of Computer Assisted Learning","volume":"42 1","pages":""},"PeriodicalIF":4.6,"publicationDate":"2025-12-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145695216","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"教育学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Automatic Short-Answer Grading in Sustainability Education: AI–Human Agreement 可持续教育中的自动简答评分：AI-Human协议

IF 4.6 2区教育学 Q1 EDUCATION & EDUCATIONAL RESEARCH

Journal of Computer Assisted Learning

Pub Date : 2025-12-02 DOI: 10.1002/jcal.70160

Emrah Emirtekin, Yasin Özarslan

Background

Sustainability education emphasises critical thinking and interdisciplinary understanding, making the assessment of students' learning outcomes complex. While Large Language Models (LLMs) have shown promise in educational assessment, their reliability in domains requiring contextual reasoning—such as sustainability—remains unclear.

Objectives

This study aims to evaluate the agreement between human raters and several LLMs (GPT-4o, Gemini 2.0 Flash, DeepSeek V3, LLaMA 3.3) in assessing short-answer responses from a university-level Sustainability course. It also investigates how this agreement varies across cognitive skill levels.

Methods

A total of 232 short-answer responses were evaluated using a rubric aligned with Bloom's Revised Taxonomy. Consensus scores from human raters were compared to LLM-generated scores using multiple statistical measures, including Quadratic Weighted Kappa (QWK), Intraclass Correlation Coefficient (ICC), Pearson correlation, and distributional overlap.

Results

Moderate agreement was found between LLMs and human raters in total scores (QWK: 0.585–0.640; r: 0.660–0.668; $� � � \overset{η}{�} � � �$ : 0.681–0.803). Inter-rater reliability among humans was good to excellent (ICC: 0.667–0.800). Criterion-level agreement declined as cognitive complexity increased, with notably low agreement on evaluating higher-order skills.

Conclusions

Overall, LLM–human agreement was moderate on total scores but declined at higher cognitive levels, indicating that LLMs are suitable for basic comprehension checks while human oversight remains necessary for complex reasoning.

可持续发展教育强调批判性思维和跨学科理解，使学生学习成果的评估变得复杂。虽然大型语言模型（llm）在教育评估中表现出了希望，但它们在需要上下文推理的领域（如可持续性）的可靠性仍不清楚。本研究旨在评估人类评分者与几个法学硕士（gpt - 40、Gemini 2.0 Flash、DeepSeek V3、LLaMA 3.3）在评估大学水平的可持续性课程的简短回答时的一致性。它还调查了这种共识在不同认知技能水平上的差异。方法采用Bloom修订分类法对232份简短回答进行评估。使用多种统计方法，包括二次加权Kappa （QWK）、类内相关系数（ICC）、Pearson相关性和分布重叠，将人类评分者的共识分数与llm生成的分数进行比较。结果llm与人类评分者在总分上有中等程度的一致性（QWK: 0.585-0.640; r: 0.660-0.668; η′$$ hat{eta} $$: 0.681-0.803）。人与人之间的信度从好到优（ICC: 0.667-0.800）。标准水平的一致性随着认知复杂性的增加而下降，在评估高阶技能方面的一致性明显较低。总体而言，法学硕士与人类的一致性在总分上是中等的，但在更高的认知水平上有所下降，这表明法学硕士适用于基本的理解检查，而人类的监督对于复杂的推理仍然是必要的。

{"title":"Automatic Short-Answer Grading in Sustainability Education: AI–Human Agreement","authors":"Emrah Emirtekin, Yasin Özarslan","doi":"10.1002/jcal.70160","DOIUrl":"https://doi.org/10.1002/jcal.70160","url":null,"abstract":"<div>\u0000 \u0000 \u0000 <section>\u0000 \u0000 <h3> Background</h3>\u0000 \u0000 <p>Sustainability education emphasises critical thinking and interdisciplinary understanding, making the assessment of students' learning outcomes complex. While Large Language Models (LLMs) have shown promise in educational assessment, their reliability in domains requiring contextual reasoning—such as sustainability—remains unclear.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Objectives</h3>\u0000 \u0000 <p>This study aims to evaluate the agreement between human raters and several LLMs (GPT-4o, Gemini 2.0 Flash, DeepSeek V3, LLaMA 3.3) in assessing short-answer responses from a university-level Sustainability course. It also investigates how this agreement varies across cognitive skill levels.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Methods</h3>\u0000 \u0000 <p>A total of 232 short-answer responses were evaluated using a rubric aligned with Bloom's Revised Taxonomy. Consensus scores from human raters were compared to LLM-generated scores using multiple statistical measures, including Quadratic Weighted Kappa (QWK), Intraclass Correlation Coefficient (ICC), Pearson correlation, and distributional overlap.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Results</h3>\u0000 \u0000 <p>Moderate agreement was found between LLMs and human raters in total scores (QWK: 0.585–0.640; <i>r</i>: 0.660–0.668; <span></span><math>\u0000 <semantics>\u0000 <mrow>\u0000 <mover>\u0000 <mi>η</mi>\u0000 <mo>̂</mo>\u0000 </mover>\u0000 </mrow>\u0000 <annotation>$$ hat{eta} $$</annotation>\u0000 </semantics></math>: 0.681–0.803). Inter-rater reliability among humans was good to excellent (ICC: 0.667–0.800). Criterion-level agreement declined as cognitive complexity increased, with notably low agreement on evaluating higher-order skills.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Conclusions</h3>\u0000 \u0000 <p>Overall, LLM–human agreement was moderate on total scores but declined at higher cognitive levels, indicating that LLMs are suitable for basic comprehension checks while human oversight remains necessary for complex reasoning.</p>\u0000 </section>\u0000 </div>","PeriodicalId":48071,"journal":{"name":"Journal of Computer Assisted Learning","volume":"42 1","pages":""},"PeriodicalIF":4.6,"publicationDate":"2025-12-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145695161","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"教育学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Who Gives Feedback Matters: Student Biases Towards Human and AI-Generated Formative Feedback 谁给予反馈重要：学生对人类和人工智能生成的形成性反馈的偏见

IF 4.6 2区教育学 Q1 EDUCATION & EDUCATIONAL RESEARCH

Journal of Computer Assisted Learning

Pub Date : 2025-12-02 DOI: 10.1111/jcal.70153

Tanya Nazaretsky, Paola Mejia-Domenzain, Vinitra Swamy, Jibril Frej, Tanja Käser

Background

Feedback is essential for learning, helping individuals understand and improve their performance. However, providing timely, personalised feedback in higher education is challenging. Generative AI offers a scalable solution, yet little is known about students' biases towards AI-generated feedback.

Objectives

This study aims to investigate how the identity of the feedback provider (human vs. AI) affects students' perceptions of feedback quality and credibility.

Methods

The study involved 472 students across diverse academic programmes and levels in authentic educational environments and employed a within-subject experimental design with a priming effect. A mixed-methods approach combined quantitative analysis of feedback evaluations with qualitative insights into students' perceptions to deepen understanding of the observed biases.

Results and Conclusions

Students perceived AI as a significantly less credible feedback provider and tended to associate lower feedback quality with AI. Disclosing the feedback provider's identity led to decreased evaluations of AI-generated feedback and an increased preference for human-crafted feedback. These patterns were consistent across academic levels, genders, and fields of study. These insights highlight the need for targeted interventions, such as improving AI literacy and building human-in-the-loop systems, to mitigate biases and enhance the effectiveness of AI in educational feedback systems.

反馈对于学习是必不可少的，它能帮助个人理解并提高他们的表现。然而，在高等教育中提供及时、个性化的反馈是一项挑战。生成式人工智能提供了一个可扩展的解决方案，但我们对学生对人工智能生成的反馈的偏见知之甚少。本研究旨在探讨反馈提供者的身份（人类与人工智能）如何影响学生对反馈质量和可信度的看法。方法在真实的教育环境中对472名不同专业和水平的学生进行了研究，采用了具有启动效应的受试者内实验设计。混合方法方法结合了反馈评估的定量分析和对学生看法的定性见解，以加深对观察到的偏见的理解。结果和结论学生认为人工智能是一个可信度明显较低的反馈提供者，并倾向于将较低的反馈质量与人工智能联系起来。披露反馈提供者的身份导致对人工智能生成的反馈的评估降低，并增加了对人工制作的反馈的偏好。这些模式在学术水平、性别和研究领域都是一致的。这些见解强调需要有针对性的干预措施，例如提高人工智能素养和建立人在循环系统，以减轻偏见并提高人工智能在教育反馈系统中的有效性。

{"title":"Who Gives Feedback Matters: Student Biases Towards Human and AI-Generated Formative Feedback","authors":"Tanya Nazaretsky, Paola Mejia-Domenzain, Vinitra Swamy, Jibril Frej, Tanja Käser","doi":"10.1111/jcal.70153","DOIUrl":"https://doi.org/10.1111/jcal.70153","url":null,"abstract":"<div>\u0000 \u0000 \u0000 <section>\u0000 \u0000 <h3> Background</h3>\u0000 \u0000 <p>Feedback is essential for learning, helping individuals understand and improve their performance. However, providing timely, personalised feedback in higher education is challenging. Generative AI offers a scalable solution, yet little is known about students' biases towards AI-generated feedback.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Objectives</h3>\u0000 \u0000 <p>This study aims to investigate how the identity of the feedback provider (human vs. AI) affects students' perceptions of feedback quality and credibility.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Methods</h3>\u0000 \u0000 <p>The study involved 472 students across diverse academic programmes and levels in authentic educational environments and employed a within-subject experimental design with a priming effect. A mixed-methods approach combined quantitative analysis of feedback evaluations with qualitative insights into students' perceptions to deepen understanding of the observed biases.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Results and Conclusions</h3>\u0000 \u0000 <p>Students perceived AI as a significantly less credible feedback provider and tended to associate lower feedback quality with AI. Disclosing the feedback provider's identity led to decreased evaluations of AI-generated feedback and an increased preference for human-crafted feedback. These patterns were consistent across academic levels, genders, and fields of study. These insights highlight the need for targeted interventions, such as improving AI literacy and building human-in-the-loop systems, to mitigate biases and enhance the effectiveness of AI in educational feedback systems.</p>\u0000 </section>\u0000 </div>","PeriodicalId":48071,"journal":{"name":"Journal of Computer Assisted Learning","volume":"42 1","pages":""},"PeriodicalIF":4.6,"publicationDate":"2025-12-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://onlinelibrary.wiley.com/doi/epdf/10.1111/jcal.70153","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145695162","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"教育学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

A Meta-Analysis of Moderators of the Effects of Technology-Enhanced Adaptive Learning on Primary and Secondary Students' Learning Outcomes 技术增强适应性学习对中小学生学习成果影响因素的meta分析

IF 4.6 2区教育学 Q1 EDUCATION & EDUCATIONAL RESEARCH

Journal of Computer Assisted Learning

Pub Date : 2025-12-02 DOI: 10.1002/jcal.70168

Jingxian Wang, Dineke E. H. Tigelaar, Tong Ye, Wilfried Admiraal

Background

Technology-enhanced adaptive learning, which includes Intelligent Tutoring Systems (ITSs), represents a significant research stream targeted at individualised learning. Although systematic review and meta-analysis studies have synthesised the evidence of technology-enhanced adaptive learning, most did not differentiate how this effect may promote different learning outcomes to varying degrees and especially did not shed much light on potential moderating effects.

Objectives

This comprehensive meta-analysis (CMA) is focused on the effects of technology-enhanced adaptive learning on cognitive, affective, and behavioural learning outcomes, considering a series of specific moderator variables.

Methods

This meta-analysis synthesised 69 studies from electronic databases and high-impact journals in educational technology published between 2012 and 2021, with a total of 9095 students.

Results and Conclusions

The findings indicate that technology-enhanced adaptive learning has a medium-positive effect on primary and secondary school students' learning in terms of cognitive, affective, and behavioural learning outcomes. Further moderator analyses revealed that factors in categories of research context, adaptive strategy, technology, and study quality significantly moderated the summary effect sizes for at least one learning outcome dimension. Twelve significant moderators for cognitive learning outcomes (e.g., subjects, adaptive targets, adaptive technology), two for affective learning outcomes (i.e., feedback type, data collection methods), and two for behavioural learning outcomes (i.e., learner characteristics, hardware used in experimental groups) were identified. The implications of the findings and directions for technology-enhanced adaptive learning practices are discussed.

Takeaway

Stakeholders need to be aware of what matters in technology-enhanced adaptive learning; it is not just the effectiveness of these adaptive learning systems but knowing how to develop, design, implement, and evaluate the effects that matter.

技术增强的适应性学习，包括智能辅导系统（ITSs），代表了针对个性化学习的重要研究流。尽管系统回顾和荟萃分析研究已经综合了技术增强适应性学习的证据，但大多数研究并没有区分这种影响如何在不同程度上促进不同的学习结果，特别是没有阐明潜在的调节效应。本综合荟萃分析（CMA）着眼于技术增强的适应性学习对认知、情感和行为学习结果的影响，考虑了一系列特定的调节变量。方法本荟萃分析综合了2012年至2021年间发表在电子数据库和高影响力教育技术期刊上的69项研究，共9095名学生。结果与结论研究结果表明，技术增强的适应性学习在认知、情感和行为学习成果方面对中小学生的学习有中等正向影响。进一步的调节分析显示，研究背景、适应策略、技术和学习质量等因素显著调节了至少一个学习结果维度的总效应大小。认知学习结果的12个显著调节因子（如受试者、自适应目标、自适应技术），情感学习结果的2个显著调节因子（如反馈类型、数据收集方法），行为学习结果的2个显著调节因子（如学习者特征、实验组中使用的硬件）。讨论了研究结果对技术增强的适应性学习实践的影响和方向。利益相关者需要意识到在技术增强的适应性学习中什么是重要的；重要的不仅仅是这些适应性学习系统的有效性，还包括知道如何开发、设计、实施和评估效果。

{"title":"A Meta-Analysis of Moderators of the Effects of Technology-Enhanced Adaptive Learning on Primary and Secondary Students' Learning Outcomes","authors":"Jingxian Wang, Dineke E. H. Tigelaar, Tong Ye, Wilfried Admiraal","doi":"10.1002/jcal.70168","DOIUrl":"https://doi.org/10.1002/jcal.70168","url":null,"abstract":"<div>\u0000 \u0000 \u0000 <section>\u0000 \u0000 <h3> Background</h3>\u0000 \u0000 <p>Technology-enhanced adaptive learning, which includes Intelligent Tutoring Systems (ITSs), represents a significant research stream targeted at individualised learning. Although systematic review and meta-analysis studies have synthesised the evidence of technology-enhanced adaptive learning, most did not differentiate how this effect may promote different learning outcomes to varying degrees and especially did not shed much light on potential moderating effects.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Objectives</h3>\u0000 \u0000 <p>This comprehensive meta-analysis (CMA) is focused on the effects of technology-enhanced adaptive learning on cognitive, affective, and behavioural learning outcomes, considering a series of specific moderator variables.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Methods</h3>\u0000 \u0000 <p>This meta-analysis synthesised 69 studies from electronic databases and high-impact journals in educational technology published between 2012 and 2021, with a total of 9095 students.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Results and Conclusions</h3>\u0000 \u0000 <p>The findings indicate that technology-enhanced adaptive learning has a medium-positive effect on primary and secondary school students' learning in terms of cognitive, affective, and behavioural learning outcomes. Further moderator analyses revealed that factors in categories of research context, adaptive strategy, technology, and study quality significantly moderated the summary effect sizes for at least one learning outcome dimension. Twelve significant moderators for cognitive learning outcomes (e.g., subjects, adaptive targets, adaptive technology), two for affective learning outcomes (i.e., feedback type, data collection methods), and two for behavioural learning outcomes (i.e., learner characteristics, hardware used in experimental groups) were identified. The implications of the findings and directions for technology-enhanced adaptive learning practices are discussed.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Takeaway</h3>\u0000 \u0000 <p>Stakeholders need to be aware of what matters in technology-enhanced adaptive learning; it is not just the effectiveness of these adaptive learning systems but knowing how to develop, design, implement, and evaluate the effects that matter.</p>\u0000 </section>\u0000 </div>","PeriodicalId":48071,"journal":{"name":"Journal of Computer Assisted Learning","volume":"42 1","pages":""},"PeriodicalIF":4.6,"publicationDate":"2025-12-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145695256","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"教育学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0