It is a challenge to promptly respond to the learning needs of students, particularly adult learners characterized by rich lived-experiences, (sometimes) negative prior learning experience, multiple and contrasting motivation types alongside considerable socio-demographic heterogeneity. Furthermore, developing a typology of adults’ learning needs remains a challenging endeavour given the different perspectives in learning needs conceptualization and the different theories concerning adult learning in both traditional and online and blended learning (OBL). In particular, no validated instrument currently exists to identify and measure adult learners' needs in OBL contexts, representing a gap that needed to be addressed. This paper developed and validated an instrument to identify adults’ learning needs based on adult learning, self-regulated and motivational theories, and technological acceptance models. We deconstructed adults’ learning needs into three categories based on the existence, relatedness, and growth (ERG) theory of human needs. Data (N=209) were collected from adult learners following blended learning programs in eight centres of adult education in Flanders, Belgium. Confirmatory factor analysis (CFA) revealed that the instrument comprising of nine dimensions of learning needs displayed adequate model fit and factor loadings. Multivariate analysis of covariance (MANCOVA) highlighted that learners who differed in educational attainment, age, and ICT skills reported differences in their needs across the sub-dimensions. Motivational orientations, however, did not significantly differentiate learning needs at the multivariate level. The validated instrument, hence, offered practitioners a diagnostic tool for needs-based OBL design. Theoretically, the study challenged the assumption that motivational orientation alone predicted differentiated need profiles in OBL, opening new avenues for debate on the interplay between motivation, prior experience, and adult learning needs.
Classifying student engagement accurately is critical for timely academic intervention; however, most existing approaches rely on arbitrarily defined thresholds that lack statistical grounding and are difficult to transfer across institutional contexts. This limitation reduces the practical applicability of engagement analytics in diverse educational settings. This study evaluates a quantile-based engagement classification framework across two contrasting datasets to assess its validity, transferability, and consistency of predictive features. Unlike threshold-based approaches, the proposed framework derives engagement categories directly from dataset-specific interaction distributions. The Open University Learning Analytics Dataset (OULAD) represents large-scale fully online learning, while the Unistudium dataset reflects a smaller blended learning context. The two datasets differ substantially in size and delivery mode, with a student ratio of approximately 17.4 to 1. This contrast provides a rigorous basis for assessing method transferability. Engagement categories (passive, moderate, and active) are derived using dataset-specific quartile thresholds (Q1 and Q3). This strategy adapts automatically to local interaction distributions and avoids manual parameter tuning. Five temporal behavioural features were extracted, including active days, unique actions, and learning consistency. Random Forest was employed as the proposed model, while a Decision Tree classifier was included as a baseline for comparative evaluation. The results indicate that the proposed framework remains effective across different educational contexts. In the OULAD dataset, the model achieved an accuracy of 92.04% with a Cohen's kappa of 0.87. In the Unistudium dataset, accuracy reached 72.50% with a Cohen's kappa of 0.59. Although performance differed between datasets, variance remained low. Feature importance analysis further revealed strong consistency across contexts, with a Spearman correlation of 0.90. Active days and unique actions were the most influential predictors in both cases. The baseline comparison further confirmed the superiority of Random Forest over the Decision Tree baseline across both datasets. These findings support e-learning practice by offering institutions a statistically grounded and automated method for engagement classification. The approach removes the need for arbitrary thresholds and reduces operational overhead in analytics deployment. From a research perspective, the study establishes realistic performance benchmarks for engagement analytics at different institutional scales, demonstrates the applicability of quantile-based engagement classification across heterogeneous datasets, and confirms that key behavioural engagement indicators transfer reliably across online and blended learning environments.
This study explores how early-years teachers evaluate and implement Computational Thinking (CT) through tangible floor-robot activities within an eTwinning Community of Practice (CoP), with attention to gender-related participation patterns. Although CT is increasingly recognised as an essential dimension of Early Childhood Education (ECE - referring to ages 4-6 according to the Greek Education System), there is still limited empirical evidence on how teachers transform CT concepts into developmentally appropriate practice and how gender may shape children’s engagement in such activities. Addressing these gaps, the study focuses on teachers’ evaluations and classroom implementation of CT, their experiences with tangible programming tasks, and their perceptions of gender-related patterns in participation, support needs, and CT performance. The research was conducted over a 24-week asynchronous professional-learning programme hosted on Moodle and involved one national cohort of Greek early-years educators (N = 473). During the professional-learning programme, participants engaged in weekly STEAM-oriented CT activities and collaboratively designed classroom learning scenarios using floor robots (e.g., Bee-Bot). Survey instruments captured teachers’ perceptions of CT, educational robotics, STEAM pedagogy, and gender-related classroom observations, while focus-group discussions provided complementary insights into classroom enactment. Following a DBR-informed exploratory mixed-methods approach, the research combined iterative engagement in the CoP with descriptive analysis of survey and focus-group data to identify patterns in teachers’ experiences and reported classroom practices. Findings indicate consistently high levels of participation for both girls and boys, with only small gender differences. Boys appeared slightly more often in the highest engagement band but were also more likely to require ongoing support, whereas girls were more frequently described as working independently and showed a modest descriptive advantage in problem solving, sequencing, and simple algorithm design. In addition, the learning scenarios were rated as useful or very useful for cultivating CT and for fostering collaboration, communication, and problem solving in early-years classrooms. Qualitative findings identified three design features as particularly effective in supporting children’s engagement: explicit sequencing supports, structured testing and debugging cycles, and the use of cooperative roles. Taken together, these findings underpin a set of practical learning-design principles for implementing CT through floor-robot activities in early childhood. More broadly, the research illustrates how an online CoP can support the adaptation of CT-focused designs into everyday classroom practice. It also contributes to e-learning research by illustrating how developmentally appropriate robotics activities within a Community of Practice may support equitable opportunities for young children to engage with foundational CT practices from the earliest years of schooling.
In recent years, medieval-themed video games have emerged as increasingly relevant educational tools, for history teaching, recognised for their ability to foster historical understanding, digital literacy, and critical thinking across a variety of learning environments. This systematic review investigates how these games are incorporated into educational practice and interrogates the narratives they construct about the medieval past. These aims are grounded in prior research on game-based learning and on the cultural analysis of medievalist representations. Guided by the PRISMA protocol, fourteen peer-reviewed studies published in the last ten years were identified, selected, and analysed to provide a structured and critical overview of current research in this area. The findings reveal a strong predominance of commercial titles particularly strategy and role-playing games that reproduce Eurocentric, militarised, and masculinised representations of the Middle Ages. Nevertheless, several studies report innovative pedagogical strategies that embed these digital resources within intentional didactic frameworks, aligning them with curricular objectives and supporting immersive experiences, enquiry-based learning, and the development of disciplinary historical competences. Such practices highlight the capacity of video games to operate as complex cultural artefacts, rather than mere motivational tools. Despite this potential, significant shortcomings remain. These include a lack of sustained critical engagement with symbolic and ideological representations, the scarce incorporation of gender-sensitive or intersectional perspectives, and the limited connection between explicit educational aims and the cultural content of the games. Addressing these gaps requires stronger pedagogical models that connect the analysis of digital representations with the development of historical thinking, digital literacy, and critical reflection. Overall, this study underscores both the opportunities and limitations of medieval-themed video games as didactic resources. It stresses the importance of inclusive, reflexive, and gender-aware approaches that challenge dominant historical imaginaries and contribute to the formation of culturally literate, critical, and democratically engaged citizens.
The development and assessment of project management competencies remain a persistent challenge in professional education. While competence frameworks emphasise behavioural and contextual judgement, assessment practices in entry-level certification contexts continue to rely predominantly on standardised knowledge-based examinations. At the same time, simulation-based and game-based learning approaches are increasingly used to support experiential learning, yet evidence linking such interventions to externally validated assessment outcomes remains limited. This study examines whether the inclusion of a simulation-based learning intervention in a preparatory course for the International Project Management Association (IPMA) Level D certification is associated with differences in certification examination performance. Using a quasi-experimental design, examination results for participants who completed simulation-supported training (n = 178) were compared with those of a reference cohort who completed the same course without the simulation component (n = 455). Certification exam scores were analysed across competence areas defined in the IPMA Individual Competence Baseline (ICB 4.0). The results indicate that participants in the simulation-supported group achieved higher overall examination scores, with statistically significant differences concentrated mainly in selected People and Practice competence elements. Mean scores in the Perspective area were also higher in the simulation-supported group, but the differences were not statistically significant. While the non-randomised design does not allow causal conclusions, the findings suggest an association between competence-oriented simulation-based learning and higher performance in a standardised, externally administered certification examination. The analysis focused not only on overall examination performance but also on the distribution of differences across individual competence elements. This made it possible to examine whether observed differences corresponded to the areas most directly activated by the simulation scenario. The study therefore provides evidence on the alignment between simulation design, competence frameworks, and externally administered assessment outcomes. The study contributes to e-learning research by demonstrating how simulation-based learning can be examined in relation to formal assessment outcomes within a professional certification context. It highlights the importance of aligning experiential learning design with competence frameworks while maintaining independence from existing assessment formats.
Papua, Indonesia, has a rich ethnochemical heritage, including plant-based dyeing and starch processing, that is rarely represented in tertiary chemistry curricula. At the same time, uneven internet connectivity constrains students’ opportunities to develop robust digital literacy. These dual challenges highlight a need for e-learning models that are culturally grounded, technologically adaptive, and pedagogically transformative. This study redesigned an Ethnochemistry Technology lecture into a low-bandwidth, culturally embedded, and virtue-infused e-learning model to improve students’ digital literacy and character outcomes measurably. Using a semester-long Design-Based Research approach at a public university in Jayapura, the course integrated Papuan case vignettes, flipped and HyFlex delivery, downloadable H5P interactives, short micro-lectures, an augmented-reality dye-extraction lab with offline alternatives, communal reflection journals, peer mentoring, and a service-learning partnership with a village craft cooperative. Cultural consultants, including community elders and a dye artisan, ensured epistemic authenticity. Learning tasks were aligned with Kurikulum Merdeka virtues, respect for local wisdom, collaboration, and environmental stewardship, and mapped to UNESCO digital-literacy domains. Findings indicate substantial and educationally meaningful gains in digital literacy and character mastery. Students demonstrated great improvement across all digital-literacy domains, alongside marked increases in demonstrated respect for local wisdom, collaborative engagement, and environmental responsibility. Engagement indicators also rose consistently throughout the semester, particularly during immersive and community-linked activities. Qualitative feedback revealed heightened confidence in evaluating digital sources and a deeper recognition of chemistry within local cultural practices. Beyond local impact, this study contributes to e-learning practice by presenting a scalable blueprint for culturally responsive, low-bandwidth digital pedagogy in resource-constrained contexts. It advances the field of e-learning in Papua by validating indigenous knowledge as a central epistemic resource in technology-enhanced higher education. The model demonstrates that digital transformation need not marginalize local culture; instead, it can amplify it. Future research should explore multi-campus replication, long-term retention effects, and adaptation for additional Indigenous language communities.
While Generative Artificial Intelligence (GenAI) allows new methods to support students in writing processes, its incorporation into university teaching needs well-designed pedagogy to ensure academic integrity and autonomy of learners, especially in Project-Based Research (PBR) courses in teacher education, where iterative feedback is important but restricted by the size of groups and limited resources. The current study aims to address the scalability problem in the feedback process by evaluating the AI-Supported Revision Module (AIRM). It is implemented as a scaffold designed according to a rubric and embedded into Moodle, using a local version of the LLaMA-2-7B model for generation of prompts based on instructor comments for revision in four modes: polishing, restructuring, justifying, and synthesizing. The purpose of the module is to provide targeted guidance to students rather than full-text generation and support revision while preserving authorship. A mixed-methods case study approach was used involving 158 primary education student teachers, out of which 132 submitted draft and revision versions, assessed using a six-criterion rubric. A mixed-design repeated measures ANOVA test revealed interaction effects for both Literature Review and Methodology at the level of p < .01, with effect sizes of d = 0.58 and d = 0.52. The results suggest more improvement in structure and synthesis for the AI group than the conventional group, while no significant differences were found for Data Justification and Interpretation, suggesting a boundary between procedural support and higher-order analytical reasoning. Process data analysis revealed active involvement of learners, evidenced by the fact that on average 14.2% of draft content was changed and 39.7% of AI suggestions were rejected. Qualitative data analysis revealed that learners utilized this AI-powered module mostly to increase text coherence and clarity while staying in control of making sense of the content, whereas teachers indicated a shift from superficial to more methodological feedback practices. Thus, the findings demonstrate that GenAI may be successfully implemented in the feedback process as an additional scaffold for revision while still respecting the agency of learners.
Online learning has become a core mode of higher education, intensifying questions about what constitutes highquality teaching and learning in digital environments and how students form evaluations of their online learning experiences. The Community of Inquiry (CoI) framework highlights teaching, social, and cognitive presence as key pedagogical conditions, yet less is known about how these conditions operate alongside learner characteristics and students' perceived learning outcomes to shape attitudes toward online learning. This quantitative study surveyed 316 undergraduates enrolled in 22 fully online undergraduate courses. Students reported demographic characteristics (including age, gender, ethnicity, faculty affiliation, learning-disability status) and self-reported cumulative Grade Point Average (GPA) range. Perceptions of CoI presences were measured using a validated CoI instrument, and perceived learning outcomes were assessed as multidimensional gains (cognitive, metacognitive, and social). Analyses included group comparisons and hierarchical regression models predicting attitudes toward online learning. Group comparisons indicated significant differences in perceived presence across demographic groups, although most effects were small. Perceptions of cognitive and social presence also varied across self-reported GPA ranges, whereas teaching presence was relatively stable. In hierarchical regression, demographic variables explained a modest portion of variance in attitudes. Adding CoI presences substantially improved prediction, with cognitive presence emerging as the primary presence-related predictor. When perceived learning outcomes were added, perceived cognitive and metacognitive gains were the strongest predictors of more positive attitudes, and the unique contribution of CoI presences was reduced, suggesting that perceived learning gains may help explain the presence-attitude link. The findings therefore point to a more integrative account of online learning quality, in which students' attitudes appear to depend less on fixed demographic differences and more on whether online courses are experienced as cognitively meaningful and supportive of reflective growth. Findings underscore the centrality of cognitive engagement and perceived learning gains for shaping students' attitudes toward online learning and point to actionable design priorities: inquiry-oriented activities, structured reflection and metacognitive scaffolds, and consistent course organization and support that promote equity across diverse learners. These results also inform institutional policy by emphasizing shared online-course quality standards and professional learning focused on evidence-informed design practices.
An open access article under CC Attribution 4.0 Abstract: This study aims to design, develop, and evaluate EsyGrade, a web-based essay grading system integrated with ChatGPT to support the assessment of conceptual understanding in economics education. This study is motivated by the prevalence of multiple-choice questions in Indonesian high schools due to efficiency considerations, while essay assessments better suited to measuring conceptual reasoning are still rarely used because of high assessment load. The study employed a simplified Research and Development (R&D) approach consisting of seven stages. Data were collected through expert validation using Aiken's V, a user response questionnaire, and pretest-post-test, and were later analysed using N-Gain and effect size (Cohen's d). The research findings indicate that EsyGrade has a high level of validity based on the expert evaluation and an excellent acceptance rate in the initial pilot test. In the main field trial, the results indicated a moderate improvement in conceptual understanding based on N-Gain, with a "large" effect size. These findings suggest that the overall change in learning outcomes indicates a fairly strong impact, although the degree of improvement varied among students. In particular, the system's key value lies in its ability to generate structured, reflective feedback, so that it serves not only as a summative assessment tool but also as a means for formative one. This study contributes to the development of e-learning by demonstrating that AI-based essay grading systems can be designed to improve efficiency while supporting deeper learning through the integration of pedagogical principles and technology.
The phenomenon of a global retirement crisis has gained increasing attention in recent years. A growing number of individuals struggle to secure sufficient financial resources to sustain themselves during retirement. Financial knowledge not only increases the tendency to save for retirement, but it also affects an individual's everyday money practices, including borrowing, saving, and investing decisions. Millennials are currently the largest cohort in the workforce and should be constantly made aware of the urgent need to save from the earliest age possible. The development and implementation of a unique gamified retirement planning application aimed at improving retirement preparedness was critically examined for its design rationale, user experiences, and observed successes and shortcomings. A randomised control experiment was conducted to observe any changes the gamified tool might have on the players' retirement preparedness. Participants, millennial employees at a higher education institution, aged between 24 and 43 were randomly divided into one of three groups (gamification group, education group, and control group), ranging between 29 and 34 participants. The gamified application created a platform where the gamification group could experience the long-term effects of their decisions in a low-risk setting by modelling real-life financial decisions, which helped them develop a more accurate grasp of retirement planning. The education group was only exposed to infographics on retirement planning, and the control group was not exposed to any intervention. Data were analysed and tested using a nonparametric ANCOVA. The gamified application tends to be more effective at improving basic financial literacy, promoting interest, fostering awareness, and building confidence. However, a decrease was observed in respondents' perceptions of how well they are prepared for retirement and whether they will achieve the income they need in retirement, which may reflect a "reality check" effect. The infographics appeared to be more effective at improving financial retirement literacy and resulted in a more positive perception of how well they are prepared for retirement, which may reflect overconfidence in their knowledge, leading them to believe they are better prepared than they actually are. Despite some content and technical limitations of the gamified application, it exhibits promise as a financial education tool, motivating individuals to take proactive steps toward retirement planning. Collectively, the evidence suggests that gamification enhances financial literacy and interest in retirement planning, prompts self-reflection, and serves as a "reality check" on respondents' perceived confidence in retirement preparedness. Consequently, e-learning environments for financial education interventions should not only consider gamification for motivation, but also balance engagement, knowledge development, and behavioural awareness, ensuring a meaningful and realistic development of financial proficiency and behaviour. The limitations of the study suggest that the sample size limits statistical power and restricts the generalisability of the findings, the self-reported measures may be subject to social desirability bias or overconfidence, particularly among participants exposed to educational material, and, lastly, due to the time limit of the study, long-term knowledge retention and sustained behavioural change could not be assessed.
An open access article under CC Attribution 4.0 Abstract: This study examines how Generative AI and knowledge-mapping tools support student learning and engagement in programming education. A quasi-experimental design was conducted with 30 undergraduate students enrolled in an object-oriented programming course, where participants used both tools across a four-week intervention. Data were collected through task performance and learner perception surveys. The results indicate that students reported higher ease of use and immediate support when using Generative AI, while knowledge mapping was associated with stronger support for conceptual understanding and reflective learning in later stages. These findings suggest that the two approaches support different aspects of learning, with Generative AI facilitating rapid clarification and knowledge-mapping tools encouraging structured conceptual engagement. The study contributes to the e-learning field by providing empirical insight into how different forms of learning support function within the same instructional context. Rather than positioning the tools as direct alternatives, the findings highlight their complementary pedagogical roles and offer guidance for integrating adaptive AI support with structured learning approaches in programming education.
This study applied the Apriori algorithm to analyze behavioral interaction patterns associated with learned helplessness (LH) in mathematics tutoring system logs. Interaction data were examined across three dimensions: LH level (low vs. high), system-based intervention (with vs. without), and problem-solving outcomes (solved vs. unsolved). The analysis of the complete dataset showed that skipping problems without using hints was the most frequent pattern linked to unsolved outcomes, while persistence behaviors such as not skipping were less dominant overall. Comparisons by LH level showed that low-LH students had stronger links between problem solving and not skipping, as well as positive associations between hint use and solved outcomes. High-LH students showed more avoidance patterns, with skipping strongly tied to unsolved outcomes. In the comparison of system-based intervention conditions, students without intervention had the highest lift for persistence-success links, while the with-intervention group had stronger patterns involving skipping behaviors leading to unsolved outcomes. Outcome-specific analysis showed that not skipping was consistently associated with solved problems across all groups, while skipping without hints predicted unsolved outcomes. Practical implications and recommendations are discussed.
Digital transformation in higher education has increased interest in faculty adoption of emerging technologies such as the Internet of Things (IoT). This study investigates faculty perceptions of IoT integration in Jordanian private universities, with particular attention to gender and academic rank. Grounded in the Technology Acceptance Model (TAM) and the Unified Theory of Acceptance and Use of Technology (UTAUT), the study examines how key acceptance constructs shape IoT adoption in teaching. A quantitative, descriptive survey design was employed using a validated 21-item questionnaire administered to 350 full-time faculty members at Al-Zaytoonah University of Jordan. The instrument demonstrated strong reliability (Cronbach's alpha = 0.91) and sound construct validity confirmed through confirmatory factor analysis (CFI = 0.95, RMSEA = 0.06). Results indicated a high overall level of acceptance of IoT applications in teaching (M = 4.12, SD = 0.88). No statistically significant differences were found by gender, while small but statistically significant differences emerged by academic rank, with assistant and associate professors reporting more positive perceptions than full professors (eta 2 = 0.024). The findings suggest that IoT acceptance is broadly shared among faculty, with academic rank functioning as a modest, context-dependent moderator. The study contributes empirical evidence on IoT-enabled e-learning practices in Middle Eastern private higher education and highlights the need for targeted professional development and institutional support strategies.
The rapid digitalization of higher education has significantly reshaped teaching and learning practices worldwide; however, the adoption of digital pedagogy among university teachers remains uneven, particularly in developing contexts such as India. This study examines the lived experiences of Indian university teachers in adopting digital pedagogy and explores the factors influencing this process within higher education institutions. Using a qualitative research design, the study employs Interpretative Phenomenological Analysis to develop an in depth understanding of how teachers perceive, experience, and make sense of digitally mediated teaching practices. Data were collected through semi structured interviews with university teachers representing diverse disciplinary backgrounds and institutional settings. The analysis followed a systematic and iterative IPA approach to identify emergent themes grounded in participants' narratives. The findings indicate that digital pedagogy adoption is shaped by a dynamic interplay of institutional, technological, and personal factors. Institutional support structures, availability of digital infrastructure, access to professional development opportunities, and collaborative peer environments emerged as key enablers of adoption. In contrast, challenges such as inadequate training, inconsistent technical support, increased workload, infrastructural disparities between institutions, and varying levels of digital confidence among teachers were identified as persistent barriers. The study further highlights the central role of teachers' beliefs, attitudes, and perceived pedagogical value of digital tools in determining the depth and sustainability of digital pedagogy integration. By foregrounding faculty perspectives, this research contributes to the limited qualitative literature on digital pedagogy adoption in Indian higher education and extends existing scholarship beyond technology acceptance oriented explanations. The study supports e learning practice by offering context specific recommendations related to faculty training, institutional policy, and digital readiness. By emphasizing teachers' lived experiences, the findings advance understanding of digital pedagogy as a socially situated and contextually embedded practice, providing a foundation for inclusive and sustainable digital transformation in higher education. The insights generated offer important implications for higher education leaders and policymakers seeking to strengthen digital capacity and enhance teaching quality in evolving educational environments.
E-learning has emerged as a cornerstone of contemporary higher education, offering flexible and technology-mediated environments that accommodate modern learning needs. Among its various modalities, blended learning (BL), which strategically integrates face-to-face and online instruction, has become a pivotal approach in higher education for enhancing learning outcomes and fostering talent cultivation. However, its successful implementation depends on the coordinated interaction of individual, technological, environmental, and course dimensions, constituting a complex network of interdependent factors that often remain fragmented in practice. Existing studies typically examine these factors in isolation and commonly rely on linear analytical approaches, providing limited insights into the systematic, comprehensive, and hierarchical understanding of the interrelationships among them. Understanding these structural interrelationships is therefore essential for identifying strategic leverage points that can optimise system performance and ensure the sustainable success of BL initiatives. To address this gap, this study proposes a systems science-based analytical framework that integrates the Grey Decision-Making Trial and Evaluation Laboratory (Grey-DEMATEL), Interpretive Structural Modelling (ISM), and Matrix Impact Cross Multiplication Applied to Classification (MICMAC). This integrated approach enables comprehensive and data-driven modelling of the causal parameters, hierarchical structure, and driving-dependence relationships among critical success factors of BL. First, ten critical success factors were identified through a systematic literature review and were then pairwise evaluated by twelve experts from a higher education institution in Thailand. Grey-DEMATEL was subsequently employed to quantify the causal properties and relative significance of these factors, while ISM was applied to construct a multi-layer hierarchical structure. MICMAC analysis further categorised the factors according to their driving and dependence powers. The results reveal a three-layer hierarchical structure of BL critical success factors, where policy support (R-C = 2.78), system quality (R-C = 1.73), and technical support (R-C = 1.62) serve as key causal drivers, forming the institutional and technological foundation of the BL system. Course design and technology experience act as mediating linkages connecting institutional mechanisms with learning outcomes, while attitude, perceived usefulness, and interaction represent outcome-level indicators of system performance. Among these factors, course design exhibits the highest level of centrality value (R + C = 18.6) with the causal structure. The findings extend the understanding of the causal hierarchy and strategic leverage points for achieving BL success, illustrate how institutional and technological investment are realised through course design to improve individual experience. The study offers actionable insights for policymakers and instructional designers to inform data-driven decision-making and strategic planning in higher education, as well as how this is implemented at the level of the individual academic.
This study explores the factors driving autonomous learning (AU) among undergraduate students in AI-enhanced education. It specifically examines the role of AI literacy (AI-L), critical thinking (CT), self-regulation (SR), and self-efficacy (SE). Data collected from Thai university students were analyzed using Structural Equation Modeling (SEM). The results show that AI-L demonstrated a strong and significant positive influence on all three mediating variables-SE ((3 = 0.99, t = 20.00), SR ((3 = 0.93, t = 18.53), and CT ((3 = 0.70, t = 7.30). SE exerted as the most powerful predictor of AU ((3 = 0.52, t = 6.38), while critical thinking had a smaller direct impact. The findings suggest that AI-L is a foundational competency that requires metacognitive support. Consequently, educators should utilize strategies like blended learning and reflective practice. These insights encourage a learner-centered approach to digital education, fostering future-ready, autonomous learners.
Although AI is being rapidly developed and applied in education, gaps remain in factors affect teachers' AI literacy. A cross-sectional survey of 1,680 teachers was conducted to explore relationships between school environment, social environment, teacher self-efficacy, and AI literacy via structural equation modeling (CFI = 0.986; RMSEA = 0.03). The results showed that teachers' AI literacy was 3.89 +/- 1 (out of 5) in total, and the theory-practice gap was significant: stronger performance in awareness (beta = 0.75) and ethics (beta = 0.76), but weaker performance in application literacy (beta = 0.72) and evaluation literacy (beta = 0.81). School environment had the strongest direct effect on AI literacy (beta = 0.270, p < 0.001), followed by teacher self-efficacy, which served as an important mediator (beta = 0.259, p < 0.001). Social environment had no direct effect on teachers' AI literacy (beta = 0.060, p = 0.362), implying that distal effects need to be mediated by school. Demographic analysis showed urban-rural differences, decline after age 40, and subject differences (science > liberal arts). Therefore, we suggest that policymakers should transfer to supporting school-level interventions with targeted resources allocation. School leaders should create supportive technological environments and self-efficacy programs. In addition, teachers should participate in hands-on training with a focus on practical skills. This study provides useful references for integrating AI into K-12 education in China.
This study extends the Theory of Planned Behavior (TPB) to explore how students' behavioral intentions toward using generative artificial intelligence (GenAI) are associated with their reflective engagement and self-directed learning (SDL) in higher education. As GenAI tools such as ChatGPT increasingly mediate learning, understanding how learners' intentions are linked to autonomous and reflective learning behaviors becomes essential. Data were collected from 149 first-year university students (predominantly female) in Vietnam who had prior experience with GenAI for academic purposes. Using Partial Least Squares Structural Equation Modeling (PLS-SEM), the study examined relationships among attitudes, subjective norms, perceived behavioral control, behavioral intention, actual use, reflection, and two dimensions of SDL, including intentional learning and self-management. The results reveal that attitudes and perceived behavioral control significantly predict students' intentions and actual use of GenAI, whereas subjective norms have no significant effect. Behavioral engagement is positively associated with reflection and both dimensions of SDL, while reflection is positively related to intentional learning and self-management, confirming its mediating role within the proposed model linking motivation-related constructs with autonomous learning outcomes. These findings highlight reflection as a metacognitive mechanism that links students' behavioral engagement with GenAI and their SDL-related outcomes. Theoretically, the study advances TPB by positioning reflection and SDL as outcome constructs within the proposed model, rather than fixed learner traits. Practically, it suggests that educators and institutions working with first-year university students or similar learner populations should integrate reflective activities and AI literacy into curricula to promote critical, ethical, and autonomous engagement with GenAI. Designing learning environments that position AI as a reflective partner, rather than merely a content generator, supports learners' self-regulation and reflective engagement. Overall, this research contributes to understanding how intentional and reflective interaction with GenAI is associated with deeper and more autonomous learning of students among first-year university students in a GenAI-supported learning context.
Amid the global push for digital transformation in higher education, there is a critical need for objective, scalable assessment tools in subjective disciplines like visual arts. Modern teacher education increasingly integrates intelligent technologies, yet the application of machine learning (ML) for formative assessment in art education remains underexplored. While ML offers scalable feedback, its capacity to evaluate subjective creativity remains contested. The study aims to examine the technical accuracy of a CNN-based model trained on a local dataset of 300 archived projects, compared to instructor evaluations, and to analyze how future teachers (N = 180) perceive algorithmic feedback in assessment contexts. A mixed-methods design was employed using a highly reliable survey instrument (Cronbach's alpha = .925) and comparative scoring analysis across four key dimensions: Technique, Composition, Color, and Creativity. Results indicate that the model aligns strongly with human assessments on technical execution (r = .426, p < .001), and moderate alignment for Composition (r = .430, p < .001) and weaker alignment for Color (r = .327, p < .001), while correlations for Creativity were notably weaker (r = .181, p = .015), indicating persistent limitations in modeling abstract artistic intent. ANOVA results revealed that students' digital literacy significantly predicts their trust in the system (F = 3.547, p = .031) and willingness to use it (F = 8.476, p < .001). Furthermore, discrepancy analysis indicated systematic divergence across proficiency levels, with the model exhibiting increasing underestimation for highly proficient students, particularly in cases involving stylistic deviation or non-standard cultural expression. The findings suggest that while the algorithm provides consistent, transparent scoring that enhances assessment literacy, it lacks the sensitivity to evaluate high-level originality due to standardization bias. This study contributes to the field by empirically demonstrating the "accuracy-creativity trade-off" in ML-based art assessment and by validating a hybrid assessment framework that balances algorithmic precision with pedagogical intuition. The study concludes that ML tools should function as "human-in-the-loop" support systems rather than autonomous graders, fostering critical reflection and digital competence in future educators.
Blended learning (BL) has become an increasingly prevalent instructional model in primary and secondary education, yet its implementation has intensified concerns about teacher workload and well-being. While prior research has documented workload pressures associated with digitalization and AI integration, there remains limited empirical insight into how teachers experience, interpret, and manage these demands in blended learning environments. Guided by the Job Demands-Resources (JD-R) model, this qualitative study investigates how BL reshapes teachers' job demands and resources, and how educators respond to these changes in practice. Specifically, this study explores how BL influences teachers' perceived workload (RQ1), the specific challenges they encounter during BL implementation (RQ2), and the strategies and resources they employ to manage these demands effectively (RQ3). Semi-structured interviews were conducted with ten primary and secondary schoolteachers in Flanders (Belgium) who had experience implementing blended learning. Data were analysed using a reflexive thematic analysis supported by NVivo, following a systematic and iterative coding process. The JDR model informed the analytical lens, enabling the identification of workload-related demands, available resources, challenges and coping strategies within teachers' everyday practice. Regarding RQ1, the findings demonstrate the dual nature of blended learning as both intensifying workload and providing supportive resources. Teachers reported increased demands from dual-mode lesson design, technological integration, expanded assessment requirements, and institutional platform mandates, leading to cognitive and emotional strain. Conversely, automated assessment, reusable digital materials, learning platforms, and inclusive tools reduced administrative effort and supported organization, partially decreasing these heightened demands. With regard to RQ2, workload pressures were intensified by challenges in digital classroom management, frequent technical disruptions, and the continuous need to learn and support new technologies. Teachers also reported emotional and organizational strain linked to ineffective collaboration, time constraints, infrastructural shortcomings, and resistance to pedagogical change, particularly during early stages of blended learning (BL) adoption. In response to RQ3, teachers described a range of coping strategies and supportive resources. These included reusing and adapting digital materials, employing AI-supported tools for automated assessment, developing digital skills through peer support, and implementing structured classroom routines. Institutional resources, such as reliable IT support, modular professional development, collaborative planning, clear BL guidelines, and leadership support, functioned as key job resources that buffered workload pressures and supported sustainable BL practices. This study contributes to the literature by applying the JD-R model to the K-12 blended learning context, offering a theoretically grounded account of how workload pressures and supports interact in teachers' daily work. Beyond documenting challenges, the findings generate actionable insights for school leaders and policymakers, highlighting the need for systemic workload-sensitive BL implementation, structured collaborative planning, and sustained professional development aligned with instructional realities. By reframing blended learning through a job demands-resources perspective, this study advances understanding of sustainable technology integration in compulsory education.
