The Psychological Assessment master’s course has undergone a substantial redesign since 2020, introducing multiple formative assessments to support active and self-regulated learning. This digitally enriched course environment enables the collection of fine-grained behavioral and performance data throughout the semester. The primary objective of this study is to develop and evaluate a robust and scalable framework that supports the provision of personalized learning recommendations using assessment data available in real time once the relevant formative assessments have been completed. Our approach uses separate principal component analyses to aggregate time-investment and performance variables into two interpretable dimensions. Median-based segmentation of the resulting component scores defines four recommendation-oriented learner categories. Although intentionally simple and transparent, the framework adapts dynamically as assessment data accumulate over the semester. Bootstrap analyses of the principal components and cross-cohort transferability tests demonstrate stable component structures and consistent prediction trajectories across three cohorts, including adjacent-year and two-year transfer scenarios. Predictive performance improves steadily as assessment data accumulate and stabilizes after the third formative assessment, administered around week 9, after which additional gains become progressively smaller. Remaining misclassifications are primarily concentrated near the median decision boundaries, consistent with the continuous distributions of time investment and performance. This enables timely mid-semester identification of recommendation-oriented categories that can inform targeted pedagogical goals related to motivation, learning strategies, and appropriate levels of challenge. Overall, the findings show that recommendation-oriented learner categories can be derived from routinely collected formative-assessment data, predicted early enough to support pedagogical intervention, and transferred consistently across cohorts. The proposed framework therefore provides a structurally robust and interpretable foundation for scalable personalized support in courses with recurring formative assessments.
Digital technologies are increasingly integrated into primary education, yet evidence remains limited regarding how digital game-based learning (DGBL) affects the relationships among musical knowledge acquisition, aural skills development, and classroom motivation. To address this limitation, this study examined the effects of DGBL in a music education context using a Cluster‑Randomized Controlled Trial. Participants were 277 third-grade students who took part in an eight-week intervention. Outcomes were estimated using Rasch‑scaled measures, and motivation measures were validated through Cronbach’s α and confirmatory factor analysis. Linear mixed models showed significant gains in musical knowledge and all six motivational factors in the intervention group, whereas a marginally significant effect was observed for aural skills, suggesting that the intervention dosage may have been insufficient. Aural skills significantly moderated the effect of DGBL on musical knowledge. We observed a possible “autonomy paradox”: aural skill gains were negatively related to empowerment, raising the possibility that scaffolded autonomy may be especially important in teacher-centered contexts such as China. The results underscore the importance of designing DGBL experiences that accommodate differential learning outcomes and suggest that DGBL implementation must be carefully adapted to local pedagogical contexts.
This study investigated university students’ perceptions and evaluations of an AI-assisted animation production learning model. By applying Semantic Network Analysis (SemNA), the study examined the linguistic expressions, conceptual structures, and underlying perspectives of students toward this emerging pedagogical approach. Participants were 158 undergraduate students aged 18–20 who had completed a course in AI-assisted animation production. Participants were selected through purposive sampling, and data were collected through semi-structured interviews guided by seven core questions. Two design-education experts independently coded the interview data using a predefined coding guide, and coding disagreements were resolved through discussion until consensus was reached. The analytical procedure included word frequency analysis, co-occurrence matrix construction, and semantic network visualization using Gephi software. Furthermore, the Louvain modularity algorithm was applied to identify semantic clusters, and network density, degree centrality, and betweenness centrality were used to analyze the network structure. The findings indicated that “Use of AI tools” was among the most prominent and recurrent central concepts across the seven semantic networks. Students generally perceived AI technologies as valuable tools for fostering creativity, streamlining animation workflows, and enhancing learning motivation and engagement. The semantic modules were primarily clustered into three domains: practical creation, affective and motivational experiences, and instructional applications. In addition, students identified several challenges, such as inconsistent visual styles and limited creative depth, and expressed expectations for stronger instructional support to integrate AI tools effectively into design education. Overall, the results revealed a generally positive student attitude toward AI-assisted animation learning, characterized by both integrative understanding and critical reflection. This study provides empirical evidence to inform curriculum development and instructional practices, while also demonstrating the value of SemNA for revealing the conceptual relationships underlying students’ reflections on their learning experiences.
A school principal’s competency-based job performance and school technology leadership skills are a key factor in school performance and the quality of the education system. In this context, an examination was conducted of school principals”” technology leadership skills and competency-based job performance, as perceived by the subjects themselves. The research was conducted in two stages. In the initial phase, an investigation was conducted into the validity and reliability of the school technology leadership scale. In the second stage, the levels of school principals”” school technology leadership skills and competency-based job performance were determined. The research comprised descriptive and correlation analyses based on quantitative research methods and validity and reliability studies. The study groups comprised 106 school principals in the validity and reliability study and 312 school principals in the study to determine their job performance levels based on school technology leadership skills and competencies. In the reliability study of the school technology leadership scale, Cronbach””s alpha and inter-item correlations were calculated, and very high values were found. Confirmatory Factor Analysis was performed for validity values, and the results were found to be at an acceptable level. The research concluded that the school technology leadership scale is a valid and reliable scale. In addition, an evaluation of the school principals”” skill levels in technology leadership revealed that these levels were moderate across the scale and in the ””Managing Applications”” sub-dimension. However, the levels were found to be high in the ””Guiding and Supporting”” sub-dimension. The job performance of school principals was found to be at a moderate level across the scale and in the sub-dimensions of ””Pedagogical Management: Assessment and Leadership in Learning and Teaching Processes”” and ””Creating Harmony and Professional Development within the School””. The remaining four sub-dimensions demonstrated similarly elevated levels of performance. As school principals perceive it, a positive, strong, and significant relationship has been identified between school technology leadership and competency-based job performance. The study is considered important in terms of proving the validity and reliability of the school technology leadership scale, determining the professional needs of school principals, and ensuring their development.
Tutoring systems enable individualized support for diverse learners. Feedback is one of the most influential factors shaping learning, yet the optimal form of feedback remains an open question. Growth mindset feedback emphasizes effort and the value of errors. It has shown promise in fostering growth mindset beliefs, academic performance, and persistence in classroom settings, though results for tutoring systems in mathematics remain mixed. Computer science may represent a particularly relevant domain for this type of feedback, given frequent errors, compiler messages, and iterative debugging, which demand both persistence and frustration tolerance. This study examined whether growth mindset feedback is effective in a tutoring system for computer science. In a controlled between-subjects experiment, 79 German secondary school students (aged 14 to 17 years) received either performance-based feedback (indicating the proportion of correct responses) or growth mindset feedback while learning code tracing. Performance, mindset beliefs, and self-efficacy were assessed pre- and post-intervention, with intrinsic motivation and coding experience included as covariates in robust regression models. No main effects of feedback type emerged on any outcome, and gender did not moderate these results. Exploratory moderation analyses, however, revealed a gender-specific pattern. Among female students with high coding experience and low intrinsic motivation, performance-based feedback was associated with positive effects on both performance and mindset beliefs, with no comparable effects for male students. These preliminary findings suggest that feedback effectiveness may depend on the interplay of gender, prior experience, and motivation, highlighting the need for future research into the underlying mechanisms.
Automated writing evaluation (AWE) systems aim to provide accurate assessment and relevant feedback regarding writing. Such tools are meant to emulate human evaluators who possess the ability to detect nuance while also alleviating human biases or inconsistencies. Importantly, user trust in these technologies is a key condition of their continued and successful use, but issues of trust are not always clearly emphasized, operationalized, or studied. In this theoretical paper, we review the utility of the concept of trust in designing and implementing systems involving human and automation interaction (HAI) and outline ways in which trust can be considered within the design and implementation of AWE using a relational trust framework (Chiou Lee, 2023). We consider varying stakeholders, their diverse goals, and barriers or aids to those goals. This paper synthesizes insights from research on trust and AWE to argue that while AWE holds promise for enhancing writing instruction and learning, effective implementation hinges on acknowledging teacher and student trust and ensuring that AWE systems are responsive and provide clear and meaningful explanations.
The integration of Generative AI into higher education has shifted the central assessment problem from whether students may use AI towards how human judgement over AI output can be preserved and evidenced. This article examines the AI as Critic mechanism within the Structured AI-Guided Education (SAGE) framework, in which students first create their own artefacts and then respond to AI review of that work. An artefact-based qualitative analysis was conducted on eight de-identified group submissions from a postgraduate Systems Analysis and Design assessment delivered across five Australian offerings, yielding forty unit-test records together with the associated AI critique logs, literature syntheses, and reflections. The analysis was organised through an Apply–Critique–Synthesise sequence. The analysed group submissions documented rejection of AI suggestions on principled grounds of scope, platform constraints, risk, compliance, and phasing rather than automatic deference to AI authority. Literature syntheses and reflections were found to contain critiques absent from AI-generated summaries, while accessibility was attended to only when explicitly cued. The findings are interpreted as evidence of what students can demonstrate under scaffolded conditions, and the limits of that evidence are made explicit. The contribution is artefact-based evidence that the AI as Critic mechanism can make human authority over AI suggestions observable and assessable, while the complementary question of how such judgement is verified as durable individual competence is left to the assurance strand of the SAGE program.
Communication, collaboration, critical thinking, and creativity—often referred to as the “4Cs”—are regarded as essential soft skills to meet the complex demands of today’s workforce. Consequently, it may be useful to facilitate the acquisition of these abilities from an early age. Meanwhile, greater attention is being given to incorporating robotics into school settings to promote useful skills. However, our understanding of how these develop through educational robotics remains limited. This experimental study aims to address this gap by investigating: (a) if tangibly programmable robots can promote the 4Cs more effectively compared to traditional methods, (b) whether differences are related to participants’ gender, age, or nature of curricular activities, and (c) the structural patterns and directional relationships among the 4Cs through path analysis. In this study, 50 preschoolers participated in 10 activities. The results showed that the experimental group outperformed the control group on the 4Cs, and gender had no significant effect on their development. Yet, it was found that age may influence communication. Additionally, the nature of activities might affect children’s skills growth. Moreover, the path analysis revealed that successful communication during robotics activities might lead to effective collaboration, which in turn may influence both critical thinking and creativity. Overall, these preliminary, exploratory findings offer valuable insights into the potential of robotics as a powerful tool for promoting 21st-century skills in early childhood education.
As financial transactions and information increasingly migrate to digital platforms, financial literacy is becoming closely connected to digital competence. This study examines how information and communication technology (ICT) conditions at the student, school, and country levels are associated with adolescents’ financial literacy performance, using data from 57,970 15-year-old students across 14 countries participating in the Programme for International Student Assessment (PISA) 2022. Using a multilevel modeling approach, the analysis situates financial literacy within digitally mediated learning environments that structure how students engage with technology-based assessment tasks across levels of the education system. The findings indicate that while individual ICT access and school-level digital policies initially show positive associations with performance, these relationships largely disappear after accounting for socioeconomic background. In contrast, students’ perceived quality of ICT access at school remains positively associated with financial literacy in the fully adjusted model, suggesting that the reliability and functionality of digital infrastructures shape how students engage with digitally mediated financial tasks. National digital readiness—measured using the Network Readiness Index—explains part of the between-country variation but operates primarily as a distal contextual factor. Overall, the findings suggest that educational inequalities in financial literacy are linked not only to digital access but also to the quality and integration of digital environments in students’ learning experiences, which serve as the foundational infrastructure for deploying future adaptive and personalized learning systems.
The rapid adoption of artificial intelligence (AI) in higher education raises a critical challenge: how can student learning be assessed when AI systems are readily available to solve complex problems? Designing AI-proof problems that reliably assess student mastery is often infeasible. To address this challenge, this study investigates a practical assessment approach implemented in an undergraduate engineering mathematics course. Interactive exercises were developed using the STACK computer algebra system and JavaScript to create personalized tasks and structured feedback within a controlled learning environment. A sequential mixed-methods design was used, combining quantitative analysis of student performance with qualitative analysis of student feedback. The results suggest that the proposed assessment method can support students’ cognitive engagement with AI while still allowing teachers to evaluate student understanding. Rather than focusing solely on AI-proof problems, the study proposes a dual pedagogical approach in which AI is integrated into assessment as a tool that supports learning while preserving meaningful evaluation of learning outcomes. The resulting framework combines interactive tasks with a hint matrix that guides students’ interaction with AI tools and supports mastery-oriented learning. Importantly, the procedure does not require the submission of sensitive or personal data to AI systems, demonstrating that AI-supported assessment can be implemented without introducing additional privacy risks in higher education contexts.
Universal design for learning (UDL) is a framework that promotes accessibility in education to accommodate the widest possible range of learners. Among UDL’s core principles, this study focuses on providing multiple means of representation, specifically through the use of summaries as an alternative way to convey the content of video lectures. We investigate the effectiveness of current AI models in generating text-based abstractive summaries as a complementary method for representing recorded lecture content. By comparing students’ perceptions of AI-generated summaries versus human-generated ones, our findings indicate that current AI models are capable of producing coherent and valuable lecture summaries. From the students’ perspective, these AI-generated summaries are seen as a beneficial addition to their learning materials. However, our study also highlights the need for educators to curate these summaries to ensure their quality. Through a quality of experience analysis, the factors shaping the learning experience and the extent of control available to learning facilitators were identified. Further research is needed to explore how lecture content influences student perceptions and to optimise AI-generated summaries for educational use.
Drawing on the Technological Frames framework, this 35-week longitudinal qualitative case study investigates how teachers in online adult education contexts perceive and respond to the implementation of gamified learning technologies over time. Through three rounds of semi-structured interviews, the study examines how teachers’ interpretations of gamification developed and identifies the conditions shaping these shifts. Findings indicate a movement from initial ambivalence, focused on teacher concerns, pedagogical legitimacy, and technical complexity, towards a more pragmatic framing of gamification as a support mechanism for promoting student self-directed learning. At the same time, teachers reported persistent frictions, including the paradox that increased student engagement could also heighten demands on their own workload. The study contributes a temporally rich and ecologically grounded account of teachers’ perceptions of educational gamification, underscoring the interpretive processes through which they frame and reframe gamified technologies. It further highlights the organisational conditions that facilitate or constrain teachers’ endorsement. These insights are relevant for scholars and practitioners interested in the strategic use of gamification in adult learning environments.
The implementation of technology in teaching brings benefits but also presents physical, psychological, technical, and ethical challenges. Therefore, a balanced approach to technology use has become highly relevant, with teacher self-reflection recognized as a key driver in fostering digital well-being (DWB). Guided by Self-Determination Theory (SDT), this study evaluates how a dedicated tool supports teachers in reflecting on their DWB. Data were collected from a digital competence training workshop involving Catalonian educators (n = 46) through the tool’s application and pre- and post-questionnaires, and examined using reflection and thematic analysis. Following the intervention, educators demonstrated deeper reflection in their responses, covering a broader range of topics. Competence emerged as the most frustrated and reflected upon need, while relatedness was the least frustrated and satisfied. This paper contributes to SDT-based research on teacher DWB through reflective practices and offers practical implications and future research directions.
Engineering drawing is a core subject in vocational education, yet many pre-service vocational teachers face challenges integrating computer-aided design (CAD) tools and emerging technologies into instruction. Generative AI (GenAI) offers new opportunities for creating adaptive learning content and visual representations, but little is known about teachers’ perceptions of its benefits in engineering drawing education. This study examined pre-service vocational teachers’ perceptions of GenAI using the Expectancy-Value Theory (EVT) framework, focusing on perceived benefit of AI (BOA), knowledge of AI (KOA), value of AI (VOA), and cost of AI (COA). Data were collected from 266 pre-service vocational teachers through a validated questionnaire administered after a hybrid Vocational Teacher Development Program (VTDP) integrating CAD, 3D modeling, and ChatGPT-based instructional design tasks. Data analysis comprised four key procedures. Descriptive statistics were used to summarize participant demographics, while independent t-tests assessed differences in perceptions across educational backgrounds. Confirmatory factor analysis was conducted to establish the validity and reliability of the measurement model. Finally, structural equation modeling was employed to examine the hypothesized relationships and mediation effects among the principal constructs. Results indicated no significant differences between participants with vocational and general high-school backgrounds across all constructs. SEM revealed that BOA significantly predicted KOA, VOA, and COA, with KOA partially mediating the relationship between BOA and VOA. These findings highlight the central role of benefit perception in shaping GenAI adoption readiness. Teacher-education programs should embed benefit-driven demonstrations and AI-literacy activities to foster effective and sustainable GenAI integration in vocational training.
Teaching materials play an important role in GIS education, where learning outcomes depend on practical skills and hands-on experience. Modern GIS instruction therefore requires materials that combine theoretical explanations with practical guidance, including step-by-step tutorials, video manuals, and screencasts. This study examines geography students’ perceptions of hybrid GIS teaching materials at two Slovak universities. The research focuses on the materials Geographic Information Systems—Creation of Selected Thematic Maps (2021) and is based on a questionnaire survey of 63 students. Data were analysed using the semantic differential method, comparing responses between institutions and between scientific and teacher-training programmes. The results show only minor differences between universities, mainly related to perceived usefulness and the use of similar materials in other courses. Scientifically oriented students evaluated the materials more positively due to more frequent GIS use. Video tutorials and step-by-step guides were rated as the most beneficial for understanding and independent work.
The rise of generative AI (GenAI) has challenged educators to rethink writing instruction and academic integrity in higher education. While some students use GenAI to support idea generation and organization, others are hesitant due to ethical or practical concerns. This study examined how students perceive the appropriateness of using GenAI during the writing process and which factors predict self-reported GenAI use. Drawing on self regulated writing theory, the study assessed writing practices associated with three phases (i.e., planning, monitoring, and reviewing), along with perceived usefulness and weekly writing behaviors. Eighty-one graduate students completed a cross-sectional questionnaire measuring their beliefs about the appropriateness of GenAI-assisted writing practices, perceived usefulness of GenAI tools, and self-reported writing behaviors. Descriptive statistics, correlation analyses, and hierarchical multiple regression revealed that students generally viewed GenAI-supported practices as appropriate, particularly during monitoring-related writing activities. After controlling for demographic and writing-related background variables, perceived usefulness and monitoring-related AI-assisted writing practices emerged as significant positive predictors of self-reported GenAI use. The final model explained 45.5
Generative artificial intelligence (GenAI) in K-12 education is a topic of interest and debate for both teachers and researchers, since its use requires the reimagining of teachers’ knowledge (Mishra et al., 2023). The goal of this exploratory study is to analyse teachers’ considerations and, using the TPACK framework, elucidate the knowledge they used when considering the place of GenAI in K-12 education. This paper draws on responses to an online questionnaire by a convenience sample of N = 153 teachers. The responses were coded by the authors using TPACK to determine the knowledge teachers employed when using GenAI for learning and teaching tasks. According to these teachers, GenAI offers a range of opportunities in the areas of teaching and learning, and planning and preparation. However, the effective use of GenAI requires specific skills and critical thinking. They also report that GenAI creates challenges, including issues related to assessment and academic integrity, relevance as an information source, and the consequences of cognitive offloading. Analysis of the teachers’ TPACK indicates that they used TK, XK, TPK, and TCK when considering GenAI in K-12 education, providing empirical evidence that advances Mishra et al.’s theoretical propositions.
There are currently few established scales for evaluating learning progression in generative artificial intelligence (GenAI). This study validates an instrument designed to assess the progression of the GenAI concept using a diverse cohort of 635 participants in Hong Kong, including 355 adolescents (under 18) and 280 adults (over 18), comprising students, teachers, parents, and administrative staff. The findings indicate that a three-parameter logistic model provides the best fit for the instrument when compared to other Item Response Theory (IRT) models. Analysis of the difficulty, discrimination, and pseudo-guessing parameter confirms that the instrument effectively differentiates individuals with various abilities. The study’s contribution is the validated Generative Artificial Intelligence Concept Test—a practical and sensitive tool applicable for assessing GenAI learning achievements across both adolescents and adults. To evaluate educational impact, a paired statistical analysis comparing performance before and after a 30-h GenAI literacy course revealed significant learning progression because of the intervention. Both groups demonstrated notable improvements, with adults showing greater improvement than adolescents. The findings highlight an urgent need to unleash the potential of adults and prioritise GenAI literacy by deepening their understanding and application of these technologies.
Despite many pedagogical benefits, peer assessment is not commonly used due to educators’ concerns about its “objectivity”. This mixed-methods study examines the quality of peer assessment in higher education and contributes to peer-assessment research by distinguishing between factual and interpretive indicators in assessment criteria. Conducted among 168 graduate students in an online course, the study focused on a task requiring students to prepare digital books. These artifacts were assessed by peers and later compared to evaluations by an expert assessor. Students also reflected on the peer-assessment process, providing qualitative insights. The findings revealed differences in assessment quality. For factual indicators, such as completeness of required content, peer assessments tended to be more generous than the expert’s evaluations. However, for interpretive indicators, such as the contribution of the book to the reader’s understanding, peer assessments were generally more critical than the expert assessor. These results suggest that, although peer assessments aligned overall with expert evaluations, reliability varied by the nature of the assessment criteria. Student reflections highlighted significant learning benefits from engaging in peer assessment. Students appreciated constructive feedback from peers, which they described as beneficial for improving their work. Many also reported using the peer-assessment criteria as a framework for analyzing the feedback they received. This internalization of the assessment process underscores the pedagogical value of peer assessment beyond task performance. The findings show that differences between peer and expert evaluations are not uniform but depend on the type of indicator assessed, offering guidance for rubric design and peer-assessment practice.
As screen-vision multimodal AI tools increasingly enter higher-education learning workflows, empirical evidence about their pedagogical impact—and potential risks for teaching and learning processes—remains scarce. This study examines the potential impact of Large Multimodal Language Models (MLMs) with screen-vision capabilities on students’ performance and perceptions during controlled learning activities among master’s and undergraduate students at University of Alicante. A quasi-experimental mixed-methods design was employed. Two cohorts were established: a control group using Gemini (an MLM without screen-vision) and an experimental group using Google AI Studio (with screen-vision). Perceptions were assessed via the Technology Acceptance Model (TAM) questionnaire, alongside pre- and post-tests, assessment rubrics, and a focus group. Results provide evidence of task-dependent benefits associated with screen-vision, with the clearest between-group differences observed in organization-related activities. Students generally evaluated Google AI Studio positively, although technical limitations and ethical concerns were also reported. The study discusses implications for instructional practice and offers recommendations for integrating screen-vision AI in higher education.
