Pub Date : 2025-01-02DOI: 10.1016/j.caeai.2025.100361
Yuk Mui Elly Heung , Thomas K.F. Chiu
Generative artificial intelligence, such as ChatGPT, has been increasingly integrated into education to change student learning experience. Current empirical studies have mixed results on how ChatGPT impacts student behavioral, cognitive, and emotional engagement. This systematic literature review and meta-analysis explores whether and how ChatGPT impacts student behavioral, cognitive, and emotional engagement. We used the PRISMA method to select, analyze, and report the results. We screened 766 articles from four databases and identified 17 empirical studies with 1735 students for analysis. We compared the effect on student engagement between ChatGPT-based and non-ChatGPT learning. We found a medium effect size on overall student engagement in ChatGPT-based learning in the random effects model. Our analyses further suggest that ChatGPT-based learning is more effective in fostering student behavioral (medium effective size), cognitive (large effective size), and emotional engagement (medium effective size) than non-ChatGPT learning. Our findings revealed ChatGPT is an effective tool for engaging students in learning. We also suggested three roles ChatGPT plays in fostering student engagement: personalized tutoring, programming and technical assistance, and content generation and collaboration. Our systematic literature review revealed potential risks and results in student disengagement, such as over-reliance.
{"title":"How ChatGPT impacts student engagement from a systematic review and meta-analysis study","authors":"Yuk Mui Elly Heung , Thomas K.F. Chiu","doi":"10.1016/j.caeai.2025.100361","DOIUrl":"10.1016/j.caeai.2025.100361","url":null,"abstract":"<div><div>Generative artificial intelligence, such as ChatGPT, has been increasingly integrated into education to change student learning experience. Current empirical studies have mixed results on how ChatGPT impacts student behavioral, cognitive, and emotional engagement. This systematic literature review and meta-analysis explores whether and how ChatGPT impacts student behavioral, cognitive, and emotional engagement. We used the PRISMA method to select, analyze, and report the results. We screened 766 articles from four databases and identified 17 empirical studies with 1735 students for analysis. We compared the effect on student engagement between ChatGPT-based and non-ChatGPT learning. We found a medium effect size on overall student engagement in ChatGPT-based learning in the random effects model. Our analyses further suggest that ChatGPT-based learning is more effective in fostering student behavioral (medium effective size), cognitive (large effective size), and emotional engagement (medium effective size) than non-ChatGPT learning. Our findings revealed ChatGPT is an effective tool for engaging students in learning. We also suggested three roles ChatGPT plays in fostering student engagement: personalized tutoring, programming and technical assistance, and content generation and collaboration. Our systematic literature review revealed potential risks and results in student disengagement, such as over-reliance.</div></div>","PeriodicalId":34469,"journal":{"name":"Computers and Education Artificial Intelligence","volume":"8 ","pages":"Article 100361"},"PeriodicalIF":0.0,"publicationDate":"2025-01-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143145844","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-01-02DOI: 10.1016/j.caeai.2025.100362
Ying Gao
The education sector currently faces several challenges, including the subjectivity of evaluation methods, uniformity of data, and a lack of real-time feedback. This study aims to address these issues by leveraging deep learning techniques, specifically Convolutional Neural Networks (CNNs), to accurately assess and enhance the quality of university teaching. In contrast to traditional teaching quality assessment methods, which often lack rigor and comprehensiveness, this study introduces a precise and thorough evaluation framework. By integrating deep learning algorithms, the study seeks to improve the objectivity and accuracy of evaluations, facilitate personalized feedback, and foster innovation in teaching methodologies. The research process involves multiple complex stages, including data collection, preprocessing, feature extraction, model construction, training, validation, and results analysis. Multi-source data—comprising student performance data, teacher evaluations, course content, and student feedback—are used to create a robust dataset. Data encoding, standardization, and feature engineering techniques are employed to enhance model input. Experimental results demonstrate that the CNN model achieves prediction accuracies of 92% for “Excellent,” 88% for “Good,” 85% for “Average,” and 80% for “Poor” in the test set. These results underscore the model's high performance in classification tasks, particularly in accurately identifying high-quality teaching, with both high precision and recall. This study not only addresses a gap in the field by utilizing multi-source data for comprehensive evaluation but also validates the effectiveness of deep learning models in assessing teaching quality. Additionally, the study provides a foundation for developing targeted teaching improvement strategies.
{"title":"Deep learning-based strategies for evaluating and enhancing university teaching quality","authors":"Ying Gao","doi":"10.1016/j.caeai.2025.100362","DOIUrl":"10.1016/j.caeai.2025.100362","url":null,"abstract":"<div><div>The education sector currently faces several challenges, including the subjectivity of evaluation methods, uniformity of data, and a lack of real-time feedback. This study aims to address these issues by leveraging deep learning techniques, specifically Convolutional Neural Networks (CNNs), to accurately assess and enhance the quality of university teaching. In contrast to traditional teaching quality assessment methods, which often lack rigor and comprehensiveness, this study introduces a precise and thorough evaluation framework. By integrating deep learning algorithms, the study seeks to improve the objectivity and accuracy of evaluations, facilitate personalized feedback, and foster innovation in teaching methodologies. The research process involves multiple complex stages, including data collection, preprocessing, feature extraction, model construction, training, validation, and results analysis. Multi-source data—comprising student performance data, teacher evaluations, course content, and student feedback—are used to create a robust dataset. Data encoding, standardization, and feature engineering techniques are employed to enhance model input. Experimental results demonstrate that the CNN model achieves prediction accuracies of 92% for “Excellent,” 88% for “Good,” 85% for “Average,” and 80% for “Poor” in the test set. These results underscore the model's high performance in classification tasks, particularly in accurately identifying high-quality teaching, with both high precision and recall. This study not only addresses a gap in the field by utilizing multi-source data for comprehensive evaluation but also validates the effectiveness of deep learning models in assessing teaching quality. Additionally, the study provides a foundation for developing targeted teaching improvement strategies.</div></div>","PeriodicalId":34469,"journal":{"name":"Computers and Education Artificial Intelligence","volume":"8 ","pages":"Article 100362"},"PeriodicalIF":0.0,"publicationDate":"2025-01-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143145478","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
The rapid advances in AI technologies showed a disruptive potential in educational practices, presenting new challenges and generating new opportunities. This phenomenon has been exacerbated since the release of ChatGPT in 2022, which has permanently transformed various educational activities and sparked widespread scientific interest. Research suggests that ChatGPT can help students navigate the complexities of STEM subjects. However, only a few studies have directed attention to the use of ChatGPT in STEM subjects in upper secondary education. With the purpose of addressing this gap, the aim of this study is to explore how students experience ChatGPT for their STEM studies, encompassing their usage, perceptions, and general knowledge about this technology. Using a mixed methods approach, the data collected included a survey and semi-structured interviews involving upper secondary students. The results show widespread adoption of ChatGPT across STEM subjects among participants, particularly in biology and especially as a tool to support the understanding of concepts. Although students exhibited limited knowledge of AI, they demonstrated some effective prompting strategies to generate relevant content and tackle potential inaccuracies and hallucinations. The findings in this paper provide insights to support the exploration of students’ experiences of ChatGPT, presenting relevant topics to further research the applications of these AI technologies within STEM subjects, given their importance for future societal development.
{"title":"Exploring students’ experience of ChatGPT in STEM education","authors":"Federico Valeri, Pernilla Nilsson, Anne-Marie Cederqvist","doi":"10.1016/j.caeai.2024.100360","DOIUrl":"10.1016/j.caeai.2024.100360","url":null,"abstract":"<div><div>The rapid advances in AI technologies showed a disruptive potential in educational practices, presenting new challenges and generating new opportunities. This phenomenon has been exacerbated since the release of ChatGPT in 2022, which has permanently transformed various educational activities and sparked widespread scientific interest. Research suggests that ChatGPT can help students navigate the complexities of STEM subjects. However, only a few studies have directed attention to the use of ChatGPT in STEM subjects in upper secondary education. With the purpose of addressing this gap, the aim of this study is to explore how students experience ChatGPT for their STEM studies, encompassing their usage, perceptions, and general knowledge about this technology. Using a mixed methods approach, the data collected included a survey and semi-structured interviews involving upper secondary students. The results show widespread adoption of ChatGPT across STEM subjects among participants, particularly in biology and especially as a tool to support the understanding of concepts. Although students exhibited limited knowledge of AI, they demonstrated some effective prompting strategies to generate relevant content and tackle potential inaccuracies and hallucinations. The findings in this paper provide insights to support the exploration of students’ experiences of ChatGPT, presenting relevant topics to further research the applications of these AI technologies within STEM subjects, given their importance for future societal development.</div></div>","PeriodicalId":34469,"journal":{"name":"Computers and Education Artificial Intelligence","volume":"8 ","pages":"Article 100360"},"PeriodicalIF":0.0,"publicationDate":"2024-12-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143145452","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-12-30DOI: 10.1016/j.caeai.2024.100354
Fiammetta Caccavale , Carina L. Gargalo , Julian Kager , Steen Larsen , Krist V. Gernaey , Ulrich Krühne
Artificial Intelligence (AI) is rapidly and consistently becoming more integrated in various aspects of our lives. One of the areas where these systems are increasingly used is education. In fact, it is both being incorporated into specific curricula, allowing students the possibility to acquire skills within this field, and more recently AI has been used as a tool to facilitate the teaching and learning process. However, an increased demand and availability of these tools do not imply a successful switch from traditional to AI-supported learning.
In this work, ChatGMP, a chatbot leveraging a Large Language Model (LLM) able to conduct an interview exercise in a Master's Degree course taught at the Technical University of Denmark (DTU), is introduced. The exercise consists in a student interview of a fictitious company, represented by the teachers or ChatGMP, regarding its Good Manufacturing Practices (GMP). The aim is for the students to ask sensible and well-reasoned questions to acquire the necessary documentation to make an exhaustive report indicating whether the company is a potential fit for business. To evaluate the initiative, we compare the performance of ChatGMP to the one of the physical teachers of the course, as well as the perception of the students towards it. The results show no significant difference in the information provided by the teachers and the model, enabling the students to achieve similar learning. The students that interacted with ChatGMP are satisfied with the initiative and would likely recommend future students to perform the audit with the digital tool. This initial experiment and its positive results lay the foundation for opening the discussion on how to use LLMs in education, the opportunities they could provide, as well as their limitations and drawbacks.
{"title":"ChatGMP: A case of AI chatbots in chemical engineering education towards the automation of repetitive tasks","authors":"Fiammetta Caccavale , Carina L. Gargalo , Julian Kager , Steen Larsen , Krist V. Gernaey , Ulrich Krühne","doi":"10.1016/j.caeai.2024.100354","DOIUrl":"10.1016/j.caeai.2024.100354","url":null,"abstract":"<div><div>Artificial Intelligence (AI) is rapidly and consistently becoming more integrated in various aspects of our lives. One of the areas where these systems are increasingly used is education. In fact, it is both being incorporated into specific curricula, allowing students the possibility to acquire skills within this field, and more recently AI has been used as a tool to facilitate the teaching and learning process. However, an increased demand and availability of these tools do not imply a successful switch from traditional to AI-supported learning.</div><div>In this work, ChatGMP, a chatbot leveraging a Large Language Model (LLM) able to conduct an interview exercise in a Master's Degree course taught at the Technical University of Denmark (DTU), is introduced. The exercise consists in a student interview of a fictitious company, represented by the teachers or ChatGMP, regarding its Good Manufacturing Practices (GMP). The aim is for the students to ask sensible and well-reasoned questions to acquire the necessary documentation to make an exhaustive report indicating whether the company is a potential fit for business. To evaluate the initiative, we compare the performance of ChatGMP to the one of the physical teachers of the course, as well as the perception of the students towards it. The results show no significant difference in the information provided by the teachers and the model, enabling the students to achieve similar learning. The students that interacted with ChatGMP are satisfied with the initiative and would likely recommend future students to perform the audit with the digital tool. This initial experiment and its positive results lay the foundation for opening the discussion on how to use LLMs in education, the opportunities they could provide, as well as their limitations and drawbacks.</div></div>","PeriodicalId":34469,"journal":{"name":"Computers and Education Artificial Intelligence","volume":"8 ","pages":"Article 100354"},"PeriodicalIF":0.0,"publicationDate":"2024-12-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143145838","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-12-30DOI: 10.1016/j.caeai.2024.100353
Luca Benedetto, Gabrielle Gaudeau, Andrew Caines, Paula Buttery
Large Language Models (LLMs) can have a transformative effect on a variety of domains, including education, and it is therefore pressing to understand whether these models have knowledge of – or, in other words, how they have encoded – the specific pedagogical requirements of different educational domains, and whether they use this when performing educational tasks. In this work, we propose an approach to evaluate the knowledge – or encoding – that the LLMs have of the Common European Framework of Reference for Languages (CEFR), and use it to evaluate five modern LLMs. Our study shows that the suite of tasks we propose is quite challenging for all the LLMs, and they often provide results which are not satisfactory and would be unusable in educational applications, suggesting that – even if they encode some information about the CEFR – this knowledge is not really leveraged when performing downstream tasks.
{"title":"Assessing how accurately large language models encode and apply the common European framework of reference for languages","authors":"Luca Benedetto, Gabrielle Gaudeau, Andrew Caines, Paula Buttery","doi":"10.1016/j.caeai.2024.100353","DOIUrl":"10.1016/j.caeai.2024.100353","url":null,"abstract":"<div><div>Large Language Models (LLMs) can have a transformative effect on a variety of domains, including education, and it is therefore pressing to understand whether these models have knowledge of – or, in other words, how they have encoded – the specific pedagogical requirements of different educational domains, and whether they use this when performing educational tasks. In this work, we propose an approach to evaluate the <em>knowledge</em> – or encoding – that the LLMs have of the Common European Framework of Reference for Languages (CEFR), and use it to evaluate five modern LLMs. Our study shows that the suite of tasks we propose is quite challenging for all the LLMs, and they often provide results which are not satisfactory and would be unusable in educational applications, suggesting that – even if they encode some information about the CEFR – this <em>knowledge</em> is not really leveraged when performing downstream tasks.</div></div>","PeriodicalId":34469,"journal":{"name":"Computers and Education Artificial Intelligence","volume":"8 ","pages":"Article 100353"},"PeriodicalIF":0.0,"publicationDate":"2024-12-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143145479","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-12-30DOI: 10.1016/j.caeai.2024.100347
Abdullah Bin Nofal , Hassan Ali , Muhammad Hadi , Aizaz Ahmad , Adnan Qayyum , Aditya Johri , Ala Al-Fuqaha , Junaid Qadir
Interviewing skills play a pivotal role in the job application and search, professional development to prepare for interviewing is a neglected area of research. Professional training methods are available but are often prohibitively expensive, limiting opportunities primarily to privileged individuals. To bridge this accessibility gap and democratize access to job opportunities, there is a need to develop automated interview simulation platforms. The advent of Generative AI (GenAI) technology, in particular Large Language Models (LLMs), makes this is viable proposition but progress is hindered by the absence of open-source implementations for reproducibility and comparison, as well as the lack of suitable evaluation benchmarks and experimental setups. In particular, we do not yet know how robust such systems are and if they will be bias-free, factors that will contribute to their acceptability and use. To this end, we propose Interview Training and Education Module (ITEM), a job interview training module that combines Virtual Reality-based metaverse technology with LLM-based GenAI models. Our module creates realistic interview experiences for skill enhancement, complete with personalized feedback and improvement guidelines based on user responses. In this paper, we present an experimental evaluation of the module to ascertain its robustness, including a bias analysis. Firstly, we establish an experimental setup to gauge platform robustness by examining question similarity across varied prompts using Bidirectional and Auto-Regressive Transformers (BART) and topic modeling. Subsequently, we explore biases in three categories—country of origin, religion, and gender—by analyzing ITEM's evaluation scores while manipulating candidate backgrounds, all while keeping their responses unchanged. Our findings indicate potential biases replicated by ITEM, highlighting the need for caution in its application for personal development and training. This pioneering initiative introduces the first open-source module for job interview training within a virtual metaverse, leveraging LLM-based Generative AI, designed for extension and testing by the scientific community, thereby enhancing insights into the limitations and ethical considerations of AI-driven interview simulation platforms.
{"title":"AI-enhanced interview simulation in the metaverse: Transforming professional skills training through VR and generative conversational AI","authors":"Abdullah Bin Nofal , Hassan Ali , Muhammad Hadi , Aizaz Ahmad , Adnan Qayyum , Aditya Johri , Ala Al-Fuqaha , Junaid Qadir","doi":"10.1016/j.caeai.2024.100347","DOIUrl":"10.1016/j.caeai.2024.100347","url":null,"abstract":"<div><div>Interviewing skills play a pivotal role in the job application and search, professional development to prepare for interviewing is a neglected area of research. Professional training methods are available but are often prohibitively expensive, limiting opportunities primarily to privileged individuals. To bridge this accessibility gap and democratize access to job opportunities, there is a need to develop automated interview simulation platforms. The advent of Generative AI (GenAI) technology, in particular Large Language Models (LLMs), makes this is viable proposition but progress is hindered by the absence of open-source implementations for reproducibility and comparison, as well as the lack of suitable evaluation benchmarks and experimental setups. In particular, we do not yet know how robust such systems are and if they will be bias-free, factors that will contribute to their acceptability and use. To this end, we propose <u>I</u>nterview <u>T</u>raining and <u>E</u>ducation <u>M</u>odule (<em>ITEM</em>), a job interview training module that combines Virtual Reality-based metaverse technology with LLM-based GenAI models. Our module creates realistic interview experiences for skill enhancement, complete with personalized feedback and improvement guidelines based on user responses. In this paper, we present an experimental evaluation of the module to ascertain its robustness, including a bias analysis. Firstly, we establish an experimental setup to gauge platform robustness by examining question similarity across varied prompts using Bidirectional and Auto-Regressive Transformers (BART) and topic modeling. Subsequently, we explore biases in three categories—country of origin, religion, and gender—by analyzing <em>ITEM</em>'s evaluation scores while manipulating candidate backgrounds, all while keeping their responses unchanged. Our findings indicate potential biases replicated by <em>ITEM</em>, highlighting the need for caution in its application for personal development and training. This pioneering initiative introduces the first open-source module for job interview training within a virtual metaverse, leveraging LLM-based Generative AI, designed for extension and testing by the scientific community, thereby enhancing insights into the limitations and ethical considerations of AI-driven interview simulation platforms.</div></div>","PeriodicalId":34469,"journal":{"name":"Computers and Education Artificial Intelligence","volume":"8 ","pages":"Article 100347"},"PeriodicalIF":0.0,"publicationDate":"2024-12-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143145487","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-12-30DOI: 10.1016/j.caeai.2024.100359
Siu Cheung Kong , Jinyu Zhu , Yin Nicole Yang
Empowerment enables students to be psychologically and affectively ready to leverage the benefits of artificial intelligence (AI). However, a theory-driven scale to extend empowerment into the use of AI for problem-solving is lacking. This study developed and validated an 11-item scale of empowerment in using AI for problem-solving (EUAIPS) based on a proposed conceptual framework that synthesises empowerment and AI-related literature. The EUAIPS scale encompasses impact, self-efficacy, and meaningfulness in using AI for problem-solving. We collected data from a diverse sample of Hong Kong senior secondary and university students before (N = 477) and after the course (N = 409). Results demonstrated that the EUAIPS scale with a three-factor structure had good reliability and validity. Students also felt significantly more empowered to use AI for problem-solving after a 14-h course using AI for problem-solving. These findings empirically support the affective dimension of AI literacy and show that psychological control/competence is particularly important for students to harness AI to solve problems at the affective level. This study presents a valid instrument for researchers and practitioners to measure empowerment in using AI for problem-solving and informs AI literacy curriculum designers about including learning activities to help students realise its impact, self-efficacy, and meaningfulness.
{"title":"Developing and validating a scale of empowerment in using artificial intelligence for problem-solving for senior secondary and university students","authors":"Siu Cheung Kong , Jinyu Zhu , Yin Nicole Yang","doi":"10.1016/j.caeai.2024.100359","DOIUrl":"10.1016/j.caeai.2024.100359","url":null,"abstract":"<div><div>Empowerment enables students to be psychologically and affectively ready to leverage the benefits of artificial intelligence (AI). However, a theory-driven scale to extend empowerment into the use of AI for problem-solving is lacking. This study developed and validated an 11-item scale of empowerment in using AI for problem-solving (EUAIPS) based on a proposed conceptual framework that synthesises empowerment and AI-related literature. The EUAIPS scale encompasses impact, self-efficacy, and meaningfulness in using AI for problem-solving. We collected data from a diverse sample of Hong Kong senior secondary and university students before (N = 477) and after the course (N = 409). Results demonstrated that the EUAIPS scale with a three-factor structure had good reliability and validity. Students also felt significantly more empowered to use AI for problem-solving after a 14-h course using AI for problem-solving. These findings empirically support the affective dimension of AI literacy and show that psychological control/competence is particularly important for students to harness AI to solve problems at the affective level. This study presents a valid instrument for researchers and practitioners to measure empowerment in using AI for problem-solving and informs AI literacy curriculum designers about including learning activities to help students realise its impact, self-efficacy, and meaningfulness.</div></div>","PeriodicalId":34469,"journal":{"name":"Computers and Education Artificial Intelligence","volume":"8 ","pages":"Article 100359"},"PeriodicalIF":0.0,"publicationDate":"2024-12-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143145476","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-12-26DOI: 10.1016/j.caeai.2024.100351
Dirk Ifenthaler , Muhittin Sahin , Ivan Boo , Darshini Devi Rajasegeran , Ang Shin Yuh
{"title":"Adaptive serious games assessment: The case of the blood transfusion game in nursing education","authors":"Dirk Ifenthaler , Muhittin Sahin , Ivan Boo , Darshini Devi Rajasegeran , Ang Shin Yuh","doi":"10.1016/j.caeai.2024.100351","DOIUrl":"10.1016/j.caeai.2024.100351","url":null,"abstract":"","PeriodicalId":34469,"journal":{"name":"Computers and Education Artificial Intelligence","volume":"8 ","pages":"Article 100351"},"PeriodicalIF":0.0,"publicationDate":"2024-12-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143145477","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-12-25DOI: 10.1016/j.caeai.2024.100357
Nanda R. Jafarian, Anne-Wil Kramer
Artificial intelligence (AI) is transforming education by enabling the creation of innovative learning resources that may cater to diverse learning needs. Students with common forms of neurodiversity, such as ADHD, often face unique challenges in higher education that are not adequately addressed by current educational resources. One potentially helpful resource is audio content, which provides a flexible and accessible supplement to traditional educational materials. While audio content, such as podcasts, is widely popular, its effect on academic achievement remains underexplored. This pre-registered randomized controlled trial investigated the impact of AI-assisted audio-learning modules on academic achievement, with a particular focus on the mediating roles of motivation and reading engagement. Results showed that the audio-learning modules increased student motivation and reading engagement. Importantly, audio-learning driven increases in motivation and reading engagement boosted academic achievement. Furthermore, students with greater ADHD symptom severity particularly benefited from the audio-learning modules, as they played a crucial role in determining course success. Together, this study highlights the potential of AI-assisted audio-learning modules as a valuable tool in digital education environments, catering to diverse learning needs and improving educational outcomes.
{"title":"AI-assisted audio-learning improves academic achievement through motivation and reading engagement","authors":"Nanda R. Jafarian, Anne-Wil Kramer","doi":"10.1016/j.caeai.2024.100357","DOIUrl":"10.1016/j.caeai.2024.100357","url":null,"abstract":"<div><div>Artificial intelligence (AI) is transforming education by enabling the creation of innovative learning resources that may cater to diverse learning needs. Students with common forms of neurodiversity, such as ADHD, often face unique challenges in higher education that are not adequately addressed by current educational resources. One potentially helpful resource is audio content, which provides a flexible and accessible supplement to traditional educational materials. While audio content, such as podcasts, is widely popular, its effect on academic achievement remains underexplored. This pre-registered randomized controlled trial investigated the impact of AI-assisted audio-learning modules on academic achievement, with a particular focus on the mediating roles of motivation and reading engagement. Results showed that the audio-learning modules increased student motivation and reading engagement. Importantly, audio-learning driven increases in motivation and reading engagement boosted academic achievement. Furthermore, students with greater ADHD symptom severity particularly benefited from the audio-learning modules, as they played a crucial role in determining course success. Together, this study highlights the potential of AI-assisted audio-learning modules as a valuable tool in digital education environments, catering to diverse learning needs and improving educational outcomes.</div></div>","PeriodicalId":34469,"journal":{"name":"Computers and Education Artificial Intelligence","volume":"8 ","pages":"Article 100357"},"PeriodicalIF":0.0,"publicationDate":"2024-12-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143145483","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
The increasing number of students dropping out of school due to social, economic, personal (e.g., depression or persistent failure), and health issues is a growing concern for governments, educators, and guardians. Identifying and analyzing the factors contributing to student dropout is crucial. Various machine learning, analytical, and statistical models have been proposed to address this issue. However, the existing models have several limitations in providing a precise and automated system for predicting dropout risk and analyzing the factors behind this. Besides, generating a balanced dataset is also a limitation as ‘Dropouts’ are less than the ‘Non-dropouts’. Moreover, selecting significant features contributing to student dropout and non-dropout is also very important in developing a model. However, this study introduces a comprehensive machine learning (ML) and explainable AI (XAI) based methodology to address these limitations. Firstly, the imbalanced dataset problem was handled using the Upsampling technique by adjusting the minority class ‘Dropout’. Then, the feature selection method Recursive Feature Elimination (RFE) is used with Cross-Validation (CV) as the RFE-CV method to select the most significant features. After preprocessing, this study proposed a hybrid model named the Hybrid Logistic Regression and Neural Network (HLRNN) model, which predicts student dropout with 96% accuracy, outperforming other experimented models as well as the parent models Logistic Regression and Artificial Neural Network with 2% and 3% accuracy. Finally, the XAI model The SHapley Additive exPlanations (SHAP), and Local Interpretable Model-agnostic Explanations (LIME) are deployed to analyze the risk factors associated with student dropout. This approach aims to assist institutions and educational stakeholders in formulating policies for student retention, enabling early intervention to reduce dropout rates.
{"title":"A novel AI-driven model for student dropout risk analysis with explainable AI insights","authors":"Sumaya Mustofa, Yousuf Rayhan Emon, Sajib Bin Mamun, Shabnur Anonna Akhy, Md Taimur Ahad","doi":"10.1016/j.caeai.2024.100352","DOIUrl":"10.1016/j.caeai.2024.100352","url":null,"abstract":"<div><div>The increasing number of students dropping out of school due to social, economic, personal (e.g., depression or persistent failure), and health issues is a growing concern for governments, educators, and guardians. Identifying and analyzing the factors contributing to student dropout is crucial. Various machine learning, analytical, and statistical models have been proposed to address this issue. However, the existing models have several limitations in providing a precise and automated system for predicting dropout risk and analyzing the factors behind this. Besides, generating a balanced dataset is also a limitation as ‘Dropouts’ are less than the ‘Non-dropouts’. Moreover, selecting significant features contributing to student dropout and non-dropout is also very important in developing a model. However, this study introduces a comprehensive machine learning (ML) and explainable AI (XAI) based methodology to address these limitations. Firstly, the imbalanced dataset problem was handled using the Upsampling technique by adjusting the minority class ‘Dropout’. Then, the feature selection method Recursive Feature Elimination (RFE) is used with Cross-Validation (CV) as the RFE-CV method to select the most significant features. After preprocessing, this study proposed a hybrid model named the Hybrid Logistic Regression and Neural Network (HLRNN) model, which predicts student dropout with 96% accuracy, outperforming other experimented models as well as the parent models Logistic Regression and Artificial Neural Network with 2% and 3% accuracy. Finally, the XAI model The SHapley Additive exPlanations (SHAP), and Local Interpretable Model-agnostic Explanations (LIME) are deployed to analyze the risk factors associated with student dropout. This approach aims to assist institutions and educational stakeholders in formulating policies for student retention, enabling early intervention to reduce dropout rates.</div></div>","PeriodicalId":34469,"journal":{"name":"Computers and Education Artificial Intelligence","volume":"8 ","pages":"Article 100352"},"PeriodicalIF":0.0,"publicationDate":"2024-12-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143145481","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
扫码关注我们
求助内容：
应助结果提醒方式：
京公网安备 11010802042870号