Machine learning (ML) has emerged as a powerful tool in education that enhances data-driven decision-making and supports timely interventions. However, its application in Nigerian higher education remains underexplored, particularly where sociocultural and institutional factors may influence predictive accuracy. This study employs the CRISP-DM framework to address this gap by applying ML to predict student performance in core educational technology courses. The models were trained on a dataset of 19,961 instances collected from three Nigerian universities from 2010 to 2024. A percentage split of 80% for training and 20% for testing was adopted. Thirteen machine learning algorithms were applied to predict students’ final exam scores, which included Hist Gradient Boosting Regressor, XGBoost Regressor, and Neural Network Regressor. The Hist Gradient Boosting Regressor emerged as the best model. It achieves a Mean Absolute Error (MAE) of 7.271, Root Mean Squared Error (RMSE) of 9.309, and Median Absolute Error of 6.049. XGBoost closely followed with an MAE of 7.729 and RMSE of 9.540. The best model has been deployed on Streamlit, which is accessible at eduscore.streamlit.app, allowing educators to predict student academic performance and provide timely interventions. The study concluded that machine learning models, particularly Gradient Boosting, can effectively predict student performance and support early identification of at-risk learners in core educational technology courses. It recommends that faculties of education adopt predictive tools, such as the deployed Streamlit model, to enhance academic support and improve students’ performance.
This study investigates university students’ digital competence through a multidimensional and person‑centered analytical framework grounded in bifactor–Exploratory Structural Equation Modeling (B‑ESEM). Digital competence was conceptualized across four functional skill domains and a general competence factor, reflecting contemporary perspectives on technology‑enhanced learning and student readiness. To overcome a common limitation in traditional Latent Profile Analysis (LPA), the conflation of overall competence levels with relative skill configurations, this study employed B‑ESEM–derived factor scores to preserve both hierarchical structure and multidimensionality. Using data from 599 university students, a four‑profile solution emerged, capturing distinct patterns of digital skill strengths and weaknesses. Gender and programming experience were incorporated as covariates to better understand demographic and experiential influences on profile membership. The results highlight meaningful variability in students’ digital skill ecosystems and underscore the value of integrating hierarchical modeling with person‑centered approaches to support targeted digital literacy development. These findings offer actionable insights for designing differentiated instructional strategies and educational technology interventions aligned with learner profiles.
The objective of this study was to develop and validate a digital literacy scale grounded in the socioformative approach for university students. An instrumental, cross-sectional design was used with a sample of 174 students from Metropolitan Lima. Content validity was established through expert judgment, and construct validity was examined using exploratory factor analysis. The results supported a four-factor structure—Metacognitive-Adaptive ICT domain, Critical digital information management, Multimedia communicative production, and Digital socio-ethical practice—that explained 61% of the variance. Reliability indices were acceptable across dimensions (α=.716–.810; ω=.728–.817). The findings indicate that the scale (EADigit) is a valid and reliable tool for assessing digital literacy in higher education contexts. However, these findings should be interpreted with caution given the sociodemographic heterogeneity of the study sample with respect to gender and geographic location.
Educational assessments, from low-stakes classroom tests to high-stakes national examinations, require item pools that are valid, fair, and secure. Automated Item Generation (AIG) aims to efficiently produce large pools of calibrated test items. This paper adopts a two-part design: (1) a brief historical mapping situating LLM-based AIG within the broader AIG trajectory; and (2) a scoping review of empirical studies on LLM-based AIG for STEM assessments, published between January 2022 and January 2026. A structured search of ERIC, Lens and OpenAlex yielded 1,267 records; after deduplication and screening, 7 studies were retained for synthesis. In all studies, LLMs were primarily used to draft stems, keys, distractors, and explanations by instruction-tuned prompting, sometimes enhanced with retrieval and human-in-the-loop review. Empirical evidence on item quality is generally promising. Multiple investigations have documented acceptable expert evaluations and, in a subset of studies, psychometric properties comparable to those of human-authored items. Nevertheless, recurrent limitations have been observed, including factual inaccuracies, construct drift, low calibration of item difficulty, and variable distractor plausibility. Few studies reported robust fairness audits or provided reproducible details, such as complete prompts and decoding settings. In general, LLM-based AIG can substantially increase throughput in STEM item development, but high-stakes deployment requires layered validation protocols (expert review, pilot testing, psychometrics, and bias audits) and governance controls to ensure traceability and item security.
Programming learning, especially among informatics students, faces significant challenges related to understanding concepts and syntax, as well as low engagement in programming courses. To address these issues, this research developed DolananCoding, a serious game-based learning platform specifically designed for C++ programming learning. The platform integrates a live coding feature, enabling students to write code directly, along with a leaderboard to motivate them through healthy competition. The immersive concept in DolananCoding is designed to provide an interactive and enjoyable learning experience, aiming to enhance students’ engagement, motivation, and learning outcomes in C++ programming. This research aims to test the effectiveness of DolananCoding in supporting better and more engaging programming learning for students using three testing methods: system, unit, and user. The testing results on 276 informatics students showed that the average user satisfaction reached 4.3 out of a 5.0 scale, indicating a positive response to the DolananCoding platform.
This study investigates the relationship between principals’ transformational leadership and teachers’ digital engagement in secondary schools across Malaysia, emphasizing the influence of leadership behaviours on adopting digital technologies. Despite significant advancements in digital infrastructure across the region, disparities in effective integration remain. 114 secondary school teachers and principals in Malaysia were surveyed using a quantitative, non-experimental, correlational research design. Descriptive and regression analyses were conducted using SPSS, with results showing that transformational leadership behaviours, particularly inspirational motivation (β = .943, p < .001) and individualized consideration (β = .224, p = .027), significantly predicted higher levels of teachers’ digital engagement. However, idealized influence (β = -.053, p = .639) and intellectual stimulation (β = -.102, p = .356) did not have a significant effect. The model explained 97% of the variance in teachers’ digital engagement (R² = .970), highlighting the critical role of leadership in fostering digital adoption. The study concludes that empowering school leaders through leadership development programs can significantly enhance teachers’ engagement with digital technologies. Future research should explore longitudinal designs and include a broader sample across different educational contexts to further validate these findings.
This study examines the longitudinal impact of AI-supported writing analytics on academic literacy in multilingual higher education. Using a quasi-experimental design over two semesters with 124 undergraduates, the treatment group employed a GPT-based analytics tool, while the control group relied on conventional feedback. Using a longitudinal quasi-experimental design and repeated-measures analyses, the study found significant group effects and significant Group × Time interactions, indicating that the treatment group achieved larger and more sustained gains in academic literacy. Process-oriented indicators and learning analytics further showed a shift from surface editing to global revision, accompanied by deeper metacognitive engagement.
Changes in the global demand for labor have highlighted the need for vocational education and training (VET) systems to develop not only technical skills but also soft skills like critical thinking, communication, and adaptability. This research investigates the use of digital technologies in a flipped classroom (FC) model and their impact on the development of soft skills in vocational students with particular attention to the growing VET ecosystem in China. With data gathered from a cross-sectional quantitative survey of 270 students from a Chinese vocational college, the research seeks to assess the impact of specific digital tools, such as video lectures, simulations, chatting platforms, and collaborative workspaces, on soft skill enhancement. Regression analysis indicates the different effects of digital technologies on specific competencies, which suggests the possibility of customized tech-rich pedagogical frameworks. The research findings underscore the need to incorporate specific technology methodologies into VET teaching to align instructional practices with the realities of the workplace. Additionally, the study addresses a critically understudied area of China’s institutional and cultural VET systems, thereby broadening the scope of research literature. The study has direct relevance to educational policy, instructional design, and vocational system infrastructure planning in response to the emerging demands of the Fourth Industrial Revolution.
Video is widely used in dance appreciation courses; however, unguided viewing often leads to passive observation rather than analytical understanding. Grounded in multimedia learning theory and instructional scaffolding, this study employed a Design-Based Research (DBR) approach to develop and evaluate a teacher-guided video instructional model in a dance appreciation course in a Chinese university. The DBR process involved iterative phases of design, implementation, analysis, and refinement within an authentic classroom context. A convenience sample of 23 undergraduate dance majors participated in the study. Data were collected through a structured questionnaire consisting of Likert-scale items and open-ended responses. Quantitative data were analyzed using descriptive statistics and reliability analysis, while qualitative data were examined using thematic analysis. Findings indicated that students reported highly positive perceptions of the instructional model, with mean scores ranging from 4.00 to 4.43 on a five-point Likert scale. Qualitative results revealed improvements in observational awareness, analytical understanding, and learning engagement. The study demonstrates that structured teacher guidance can transform video viewing from passive observation into an active analytical learning process and provides a practical instructional framework for dance appreciation and other arts education contexts.
The integration of generative AI in education necessitates a clear understanding of the factors influencing its adoption by preservice teachers. This study examined the predictors of generative AI use and its impact on the academic performance of preservice teachers in Ghana, employing a quantitative cross-sectional survey design. Data were collected from 783 preservice teachers using structured questionnaires measuring task characteristics (TskC), technology characteristics (TechC), task–technology fit (TTF), attitude (ATT), subjective norms (SJN), perceived behavioural control (PBC), use of generative AI (USE), and academic performance (AP). The partial least squares structural equation modelling (PLS-SEM) analysis revealed that TskC and TechC significantly influenced TTF, which in turn influenced USE. SJN and PBC were significant predictors of USE, whereas ATT had no effect. Additionally, USE positively influenced the AP. The model explained 74.8%, 42.7%, and 14.9% of the TTF, USE, and AP variance, respectively, with strong predictive validity (Q² > 0). The findings highlight the contextual importance of task and technology alignment, social influence, and self-efficacy in shaping AI adoption, suggesting that teacher education should pay particular attention to these factors to enhance academic performance.
The rapid spread of generative artificial intelligence (GenAI) in education has intensified concerns about its impact on learners’ critical thinking, underscoring the need for reliable instruments to assess how students engage critically with AI-generated content. This study adapted, developed, and initially validated the Generative AI–Critical Thinking (GenAI–CrT) scale, an eight-item short-form instrument designed to capture EFL learners’ critical thinking in AI-enhanced learning contexts. Data from 233 Vietnamese undergraduates and graduates underwent exploratory and confirmatory factor analyses, supporting a theoretically grounded four-factor structure comprising analytical skills, logical reasoning, evidence evaluation, and open-mindedness. The scale demonstrated excellent model fit (χ²(14) = 23.49, p = .053; CFI = .990; TLI = .979; RMSEA = .054; SRMR = .023) and high internal consistency (α = .90). A regression model using a subsample (n = 189) explained approximately 7% of the variance in AI self-efficacy, suggesting that the four analytical dimensions of the GenAI–CrT function collectively, but not independently, in shaping learners’ confidence in using AI tools. Beyond psychometric validation, this study offers initial evidence that critical thinking operates as an integrated competence underpinning reflective and ethical engagement in AI-enhanced learning. The GenAI–CrT can further support teachers and curriculum designers in assessing and scaffolding learners’ analytical, evaluative, and open-minded interactions with AI-generated content.
The COVID-19 pandemic has had a significant impact, including the change in learning methods from face-to-face to non-face-to-face learning. This change in method accelerates innovation in the use of technology to support non-face-to-face learning. Immersive technologies, such as Augmented Reality (AR), Virtual Reality (VR), and Virtual laboratories, have offered authentic and interactive non-face-to-face learning experiences in science education. This study employed a systematic literature review (SLR) method to map trends and innovations in the development of immersive technology in science education from 2019 to 2024. The literature data were obtained from the Scopus database and then selected using the PRISMA method to ensure their credibility and relevance. The results indicate a significant increase in research on immersive technology. The countries with the most significant contributions are Germany and the United States, and the most prominent participants are university students and lecturers. In addition, researchers identified six categories of challenges in developing immersive technology: technical, pedagogical, psychological, evaluation, social, and content. Despite these challenges, this presents new opportunities for developing immersive technology innovations, such as cross-disciplinary innovation, adaptive and inclusive learning, and collaboration between academics and practitioners.
Virtual reality (VR) is increasingly integrated into creative education for immersive, interactive learning. While prior research confirms VR’s potential to enhance learning outcomes, its support for creativity remains underexplored. This study examines whether exploration mediates the relationship between VR use and creative performance within a project-based learning framework. A quasi-experimental study was conducted with 80 second-year visual art students, who were allocated by intact sections to either a VR–PBL group or a traditional PBL group. Over four weeks, students completed an illustration-based design project. Creative performance was rated by experts, and exploratory behavior was measured using the Experiential Learning Survey, which captures real-world problem orientation, active participation, iterative experimentation, and reflective refinement. Results showed that the VR group demonstrated significantly higher exploration than the control group; exploration was positively associated with creativity, and it partially mediated the effect of VR on creativity. These findings suggest that VR enhances creativity less by novelty than by fostering sustained exploratory learning. The study extends constructivist accounts and informs VR-based learning designs that support inquiry, iteration, and creativity in design education.
This research investigated the acceptance of the artificial intelligence (AI) art tool, Midjourney, among 970 Chinese artists and designers using the Extended Unified Theory of Acceptance and Use of Technology (UTAUT2) model and incorporating person-centered work motivation and engagement profiles. Path analysis of the entire sample showed strong explanatory power, accounting for over 60% of the variance in usage intention and behavior. Latent profile analysis of intrinsic and extrinsic motivations and engagement revealed two subgroups: the majority “Interested and Engaged”, and the minority “Disinterested and Disengaged”. As evidenced by multigroup comparison, the former exhibited significantly greater acceptance, recognizing Midjourney’s value in enhancing creative ideation, while the latter relied more on AI’s external utility for minimizing effort and providing pure hedonic satisfaction when establishing their acceptance attitudes and usage behaviors. The study advances AI-technology adoption theory in artistic contexts, introduces a novel methodology for multigroup UTAUT2 modeling with person-centered moderators, and provides practical insights for strategic collaboration between humans and AI to augment creativity.
This study aims to develop and assess the feasibility of a computational thinking skills test instrument on genetic concepts for high school students. The method used is Research & Development (R&D) with the Berkeley Evaluation and Assessment Research (BEAR) Assessment System design, which includes construct maps, item design, outcome space, and measurement models. The study was conducted in seven public high schools in Banten and the Biology Education Department of Universitas Sultan Ageng Tirtayasa in January-May 2024. The instrument’s feasibility was assessed using expert assessments (materials, evaluation, and practitioners), analysis of trial data, and student assessments via questionnaires. The average expert assessment was 98.66% (very feasible). Rasch analysis of 122 students showed item reliability of 0.5 (quite reliable) with an infit-outfit mean square in the range of 0.7-1.3, so that all questions were acceptable. Meanwhile, the average student assessment was 66.16% (practical). The results of the study show that this instrument is suitable for measuring students’ computational thinking skills in genetic concepts.
To analyze the relationship between the Scratch program and students’ computational thinking in an educational institution in Peru, relational research guidelines were followed with 384 regular elementary school students aged 6 to 12. The group had a higher percentage of males (52.6%) compared to females (47.4%). Two custom questionnaires were used to measure both variables. The study found that the aesthetic aspects of the Scratch program are the most significant component for developing computational thinking (=3.34; =2.34; =2.47), especially the dimension of computational practices. Limitations include a small sample size for analyzing sociodemographic variables, operationalization of computational thinking dimensions, and the non-causal research design. The study also emphasizes the importance of designing didactic sequence content and identifying scaffolding behaviors for acquiring new knowledge. Functionality and prior experience are the most critical and influential factors in developing computational practices.
This study aims to investigate the experiences of university students on the use of artificial intelligence (AI) for assessment in higher education in Ghana. A quantitative survey approach was employed, using an online questionnaire administered to target participants across five public universities. The findings from a final sample size of 509 indicate varying levels of exposure to AI-based tools for assessment, learning, and research. It was found that the majority of the respondents have occasionally (n=218, 42.8%) been assessed using AI-based tools, while very few have very often (n=35, 6.9%) used AI-based tools for assessment as students. It was also revealed in the study that most of the respondents used ChatGPT (n=380, 74.7%) for learning, research, and assignments. Also, AI-supported feedback systems (n=173, 34%) are the most widely used AI-based assessment method by students. Regarding pragmatic approaches to assessment in the era of AI, participants expressed general agreement towards skill-based assessments, provision of continuous feedback, integration of innovative methods, and the use of presentations, practical and application questions for written examinations, although opinions on the replacement of traditional assessments with oral presentations were more divided. These findings underscore the evolving role of AI in higher education and provide insights for educators as well as policymakers seeking to optimize assessment strategies in the era of AI. Further research is recommended to delve into the AI-based assessment methods versus traditional assessment methods in higher education.
Gamification, integrating game design elements into educational contexts, has emerged as a promising strategy to enhance student engagement, motivation, and learning outcomes. This systematic review synthesizes findings from 41 peer-reviewed studies published between 2012 and 2023, offering a comprehensive overview of the current state of gamification in education. Following PRISMA guidelines, the review evaluates the effectiveness of gamification elements such as points, badges, and leaderboards. Key Findings indicate that while gamification can foster positive learning attitudes and competition, its effectiveness depends on the context. The impact of game elements on engagement and motivation varies based on the learning environment and student characteristics. The review also highlights notable gaps in the literature, particularly around the long-term effects of gamification and the theoretical frameworks guiding its use. The implications of this research are twofold: first, it highlights effective practices for educators seeking to implement gamification strategies; second, it underscores the need for further investigation into the challenges associated with gamification, including potential drawbacks such as diminished intrinsic motivation and issues of fairness. Overall, this review contributes valuable insights for educators, policymakers, and researchers aiming to optimize gamified learning environments and enhance educational outcomes.
The preparation of competent teachers to apply information and communication technology (ICT) in teaching depends on how a student teacher interacts with ICT at their teacher’s college (TC). Tutors require instructional strategies that engage student teachers in the use of ICTs. The study employed the substitution, augmentation, modification, and redefinition (SAMR) model to evaluate the status of tutors’ ICT integration methods. A researcher collected data from 320 tutors across ten teachers’ colleges TCs. SPSS analyzed the results quantitatively through multiple linear regressions, tables, and charts. The results show that tutors integrate ICT at a low level (substitution and augmentation). The prevalence of PowerPoint over other software in multimedia content design demonstrates the lower level of ICT integration in TCs. The study recommends implementing regular and standard ICT training in multimedia and programming, teaching ICT from basic education, reviewing the education policy and ICT syllabi, adding more tutors with ICT skills, and introducing centers for educational technology in TCs to transform ICT integration from its current level to modification and redefinition.
Integrating assistive and adaptive technologies is vital for fostering inclusivity in special education. However, awareness and perceived impact of these tools among aspiring special educators remain underexplored. This study examined the awareness and perceived impact of assistive and adaptive technologies among 56 students enrolled in a Multiple Disabilities (MD) B.Ed. Special Education program. Validated scales with strong reliability and unidimensionality were employed, with the awareness scale explaining 70.62% of the variance and the perceived impact scale accounting for 53.68%. Descriptive and inferential analyses indicated that awareness levels ranged from basic to moderate (mean = 3.37), with a subset showing high awareness (mean = 4.09–4.73). Perceived impact scores were consistently positive (mean = 4.31), emphasizing the role of these technologies in fostering inclusivity. Gender- and semester-based comparisons revealed no significant differences (p > .05). A moderate positive correlation (r = .473, p < .01) between awareness and perceived impact highlights their interdependence. The findings underscore the need for targeted interventions, such as workshops and field experiences, in teacher education programs to enhance awareness and practical application. Such measures are critical for bridging gaps in teacher preparedness and promoting the effective use of assistive technologies in inclusive classrooms.
