Taejung Park

AI-based writing support systems are increasingly being explored as tools for supporting English writing in educational contexts. In English as a Foreign Language (EFL) settings, learners often need support in generating ideas, organizing content, receiving feedback, and revising written texts. However, AI-based writing tools should be carefully designed and evaluated to ensure their usability, reliability, and educational relevance before being implemented in authentic learning contexts. This study aimed to design and evaluate the educational usability of WriteOn, an AI-based writing support system for EFL writing tasks that integrates AI and natural language processing technologies, including GPT-API, NLTK, spaCy, and BERT. Although the system was initially designed to support English diary writing in EFL contexts, its feedback structure and system architecture were intended to be extensible to broader writing tasks. A formative usability evaluation was conducted with eight experts in South Korea with backgrounds in EFL instruction, educational technology, instructional design, and human–computer interaction. A structured questionnaire was used to assess the system’s effectiveness, efficiency, learnability, satisfaction, memorability, error prevention, and cognitive load minimization. Based on the expert evaluation results, the system was refined through an iterative design process. The findings provided preliminary evidence of WriteOn’s overall usability while identifying several areas for improvement, including clearer feedback, more consistent design features, and enhanced language support in grammar correction modules. In response, the system was refined to include topic and genre selection, rubric-based evaluation, and a more consistent interface. These results suggest that integrating generative AI and natural language processing technologies into EFL writing support can support the development of reliable, scalable, and learner-centered writing tools. The study also highlights the importance of iterative and formative usability evaluation in developing AI-driven writing support systems that promote instructional efficiency, feedback transparency, and learner-centered revision.

As innovation barriers and inertia work against the application of new technologies in educational settings, reducing them will help overcome innovation resistance and increase the acceptance and adoption of this technology. This study aims to examine what types of innovation barriers and inertia there are (as perceived by users), how the types of users who perceive innovation resistance could be divided, and which characteristics differ depending on the type of user in a situation where a university adopts a new LMS as a learning technology tool for innovation. This study derived risk barriers, usage/value/tradition barriers, image barriers, and inertia as four factors that affect innovation resistance. The results of the study suggest that new LMS-related personalized support and training programs should be developed according to the specific needs and characteristics of each user cluster.

KSII Transactions on Internet and Information Systems TIIS Paper Details

Collecting and analyzing log data can provide students with individualized learning to maintain their motivation and engagement in learning activities and reduce dropout in Massive Open Online Courses (MOOCs). As online learning becomes more and more important, the demand for learning analytics is surging to design a variety of interventions that can achieve learning success and achieve individual learning goals and targets. In response to significant demand, we intended to derive data standards for learning analytics by specifying more the factors influencing MOOC completion suggested in previous research results. Therefore, this study aims to compare the event logs of students who have achieved scores adjacent to the minimum passing score of Korean Massive Open Online Course (K-MOOC) completion by dividing them into the completion (C) group and the non-completion (NC) group. As a result of analyzing the log data accumulated on the 60 K-MOOCs, what is interesting in the results of this study is that there was no significant difference between the C group and the NC group in video viewing, which is considered the main learning activity on the MOOC platform. On the other hand, there was a statistically significant difference between the C group and the NC group for textbook interactions in the percentage of learners who performed and the average number of logs per learner, as well as problem interactions in the average number of logs per learner. Students’ assertive activities such as textbook interaction and problem interaction might have greater value for MOOC completion than passive activities such as video watching. Therefore, MOOC instructors and developers should explore more specific design guidelines on how to provide problems with individualized hints and feedback and offer effective digital textbooks or reference materials for the large number of students. The results suggest that collecting and analyzing MOOC students’ log data on interactions, for understanding their motivation and engagement, should be investigated to create an individualized learning environment and increase their learning persistence in completing MOOCs. Future studies should focus on investigating meaningful patterns of the event logs on learning activities in massive quantitative and qualitative data sets.

The purpose of this study is to explore the factors that have the most decisive influence on actual learning intention that leads to participation in adult education. For developing the predictive model, we used tree-based machine learning, with the longitudinal big data (2017~2020) of Korean adults. Based on the gradient boosting machine (GBM) results, among the eleven variables used, the most influential variables in predicting the possibility of lifelong education participation were self-pay education expenses and then highest level of education completed. After the grid search, not only the importance of the two variables but also the overall figures including the false positive rate improved. In future studies, it will be possible to improve the performance of the machine learning model by adjusting the hyper-parameters that can be directly set by less computational methods.

This study examines how students’ situational interest changes across different segments of peer reading activities and identifies the relationship with personal traits like the students reading skills and motivation. The authors measured their levels of situational interest, reading skills, and reading motivation, based on their participation in individual and peer reading using PALS. The result revealed that situational interest was rather stable across different segments during peer reading activities. However, the change in situational interest differed between less competent readers and more competent readers across different reading tasks. The less proficient readers showed a significant increase in situational interest before and after peer reading. Furthermore, situational interest during peer reading activities did not correlate with reading motivations, contrary to the result of the individual reading task. The findings suggest that peer reading may be more beneficial than individual reading when teaching learners with low reading skills or motivation.

This study aims to identify factors to be considered when developing Information and Communication Technology (ICT)-integrated classroom models and to suggest a conceptual framework for considering more appropriate classroom models, tailored to the environments and needs of each developing country. In-depth interviews were conducted with experts, yielding many factors to be taken into account when integrating ICT in various educational contexts of developing countries. The factors are categorized into five domains: current status and relevant policies in ICT education, ODA (Official Development Assistance) goals and priority setting, infrastructure and technical challenges, pedagogical methods, and educational resources. Based on the recommendations elicited for the five categories, we suggest a conceptual framework that provides practical guidance on the criteria for selecting classroom models suited to each partner country’s environments and contexts. This study also draws implications from the interviews for classroom design guidelines related to hardware and infrastructure, to improve the sustainability of ICT-integrated classroom projects.

The current popularity of flipped learning may be attributed to its reported successes in bringing about enhanced learner engagement and positive learning outcomes. This study aims to improve our understanding of flipped learning (FL) outcomes by examining factors of learner engagement, academic capability, and epistemological beliefs. Data were collected and statistically analyzed from 231 undergraduate students enrolled in a general biology or general chemistry course at a Korean university implemented with the FL model. The results indicated that learners’ epistemological beliefs influenced neither pre‐class nor in‐class engagement, but that academic capability did affect both pre‐class and in‐class engagement. Only content/learning‐related outcomes, not generic competencies or satisfaction, were affected by both academic capability and epistemological beliefs. Both pre‐class and in‐class engagement levels affected most types of FL outcomes except generic competencies, which were only affected by the extent of in‐class engagement. Learners’ affective engagement was not influenced by epistemological beliefs, but directly and indirectly affected FL outcomes via behavioral and cognitive engagement. Theoretical and practical implications are discussed along with recommendations.