ETR&D

This study investigates the effectiveness of two instructional strategies, instructional scaffolding (guided worksheets) versus a generative strategy (mind mapping), within a Virtual Reality (VR) music history learning environment. A quasi-experimental design was conducted with 96 junior high school students (aged 14–15 years). The intervention lasted for two weeks, consisting of two 45-min VR sessions. Mixed-methods data were collected through pre- and post-tests, motivation questionnaires, and semi-structured interviews with 11 students. Quantitative results indicated that the mind mapping group was associated with significantly better conceptual understanding compared to the guided worksheet group (p = .010, ηp2= .069) suggesting it potentially promoted deeper cognitive processing. Furthermore, the guided worksheet group experienced a significant decline in intrinsic goal orientation (p = .001, Cohen’s d = .54), whereas the mind mapping strategy successfully sustained students’ learning motivation. Regression analyses revealed distinct learning mechanisms: task adherence (Completeness) was the key predictor of success for the worksheet group, whereas structural accuracy (Correctness) predicted achievement for the mind mapping group. Qualitative findings explained these statistical differences. Qualitative themes included “blind copying” in the worksheet group and “making connections” in the mind mapping group. Interviews revealed that guided worksheets induced a split-attention effect, which increased extraneous cognitive load and diminished motivation. Conversely, mind mapping facilitated active knowledge construction, with learners noting they had to independently “catch the key points” without disruptive toggling between the physical paper and the VR headset. The study concludes that generative strategies are superior for optimizing immersive VR learning by fostering a sense of agency and transforming learners into active knowledge constructors.

This study advances a framework for understanding uncertainty in ill-structured problem-solving contexts. A systematic review and qualitative synthesis of 74 peer-reviewed articles was conducted using thematic analysis and taxonomic coding to clarify how uncertainty is conceptualized across problem-solving and learning research. The resulting framework defines uncertainty as a subjective metacognitive and psychological experience that arises as learners interpret problems, make judgments, and regulate action under conditions of incomplete, conflicting, or changing information. Seven intersecting sources of uncertainty are delineated: problem context, outcomes, sociocultural influences, epistemological factors, environmental conditions, dynamicity, and complexity. By distinguishing uncertainty from adjacent terms (for example, ambiguity and vagueness) and from its manifestations (for example, hesitation or frustration), the framework provides conceptual clarity that supports more coherent theorizing and improved operationalization in empirical work. Implications for instructional design include informing the design and facilitation of problem-based, inquiry-based, and case-based learning tasks in which uncertainty is anticipated and supported as part of complex learning.

The rapid emergence of generative artificial intelligence (GenAI) presents transformative opportunities for legal education, yet empirical research examining its pedagogical integration remains limited. This study investigates the impact of GenAI integration on legal reasoning development and learning outcomes in property law courses within Chinese higher education. Grounded in constructivist learning theory, the Technology Acceptance Model (TAM), and the Technological Pedagogical Content Knowledge (TPACK) framework, this research develops an integrated AI-Enhanced Legal Pedagogy Framework to guide effective technology integration in legal education. Employing a mixed-methods case study design, the study engaged 180 law students in a semester-long intervention using iFlytek Spark, combining quantitative assessments through pre-post legal reasoning tests and TAM-based surveys with qualitative data from interviews, reflection journals, and classroom observations. Results indicate significant associations between GenAI-enhanced instruction and improvements in legal reasoning abilities (between-group Cohen's d = 1.17), though this effect size warrants cautious interpretation given the non-randomized design and potential confounding from instructor enthusiasm and motivational differences. Technology acceptance was moderate to strong (R2 = .59 for behavioral intention). Structural equation modeling reveals that perceived usefulness (β = .52) and instructor guidance (β = .23) were associated with learning outcomes. Qualitative findings highlight GenAI's potential role as a cognitive scaffolding tool while identifying challenges in preventing over-reliance. This study offers, theoretically, a preliminary conceptual integration of existing frameworks tailored to GenAI-enhanced legal education, and, practically, evidence-based strategies for legal educators to redesign pedagogy and assessment in the AI era.

This study conducted a meta-analysis of 121 empirical studies to identify teaching- and learning-related variables that influence cognitive and affective outcomes in flipped learning environments. Teaching-related variables were classified into course design factors and instructor characteristics, while learning-related variables were classified into learning task factors and learner characteristics. The overall analysis showed that both types of variables had significant medium effects on learning outcomes (r = .352 for teaching-related variables and r = .303 for learning-related variables). For cognitive outcomes, learner-learner interaction had the highest effect size among teaching-related variables (r = .380), while learning engagement had the highest effect size among learning-related variables (r = .411). For affective outcomes, autonomy support (r = .504) and perceived usefulness (r = .394) had the highest effect sizes among teaching- and learning-related variables, respectively. Moderation analyses showed that the effects of learning-related variables increased with school level, and the effects of teaching-related variables were significantly greater when pre-class materials and in-class activities were diverse. These findings suggest that effective flipped learning depends on both the instructor’s careful instructional design and the learner’s active engagement, and that diversifying pre-class materials and in-class activities can enhance the role of teaching-related variables in promoting learning outcomes.

Instructional design skills are vital for preservice teachers to develop innovative online courses in higher education. In recent years, researchers have employed Generative AI (GAI) to cultivate these skills within inquiry-based learning frameworks. However, preservice teachers often interact with GAI solely through text, which limits their ability to engage effectively in instructional design activities due to the absence of multiplayer-GAI interaction inquiry scenarios. To address this gap, I proposed a GAI-supported scenario inquiry-based learning in multiplayer virtual reality (GAI-SIBLMVR) learning approach by integrating GAI into multiplayer virtual reality (MVR), thereby encouraging critical analysis and innovative design of online courses at each stage of inquiry-based learning. A quasi-experimental design was adopted in this study to compare the effects of GAI-SIBLMVR implemented by an experimental group, with the conventional inquiry-based learning (C-IBL) learning approach used by a control group on preservice teachers’ instructional design skills. Results indicated that the GAI-SIBLMVR group significantly outperformed the C-IBL group in both instructional design knowledge and competence scores. Furthermore, the experimental group demonstrated better performance in higher-order thinking skills, particularly in critical thinking, problem-solving, and creativity. It also revealed that preservice teachers exhibited positive learning engagement, including cognitive, behavioral, emotional, and social engagement. This study offers valuable insights for researchers and educators seeking to incorporate GAI and multiplayer virtual reality into inquiry-based learning as a means to foster instructional design skills for preservice teachers in higher education.

As generative AI makes misinformation more variable, persuasive, and context-sensitive, digital literacy education requires learning environments that allow students to practice credibility evaluation under similarly dynamic conditions while maintaining instructional control. This study examines adaptive dialogic feedback, defined as bounded AI-generated NPC responses that vary according to learners’ verification decisions and gameplay context, in an AI-enhanced serious game for misinformation evaluation. Rather than assuming that AI-driven dialogue directly improves short-term knowledge acquisition, the study investigates whether it changes motivational, experiential, and behavioral learning processes compared with scripted NPC dialogue. Grounded in self-determination theory and experiential learning, this study employed a two-condition experimental design comparing scripted NPC dialogue with AI-enhanced adaptive dialogue. Sixty undergraduate students were assigned to either an AI-enhanced NPC condition (n = 30) or a scripted NPC condition (n = 30) using gender-stratified randomization. Measures included pre–post digital literacy tests, intrinsic motivation (IMI), user engagement (UES-SF), usability (SUS), and gameplay log data. Results indicate that both conditions significantly improved digital literacy knowledge (p < .001), with no statistically significant difference in learning gains between groups (p = .066). However, the AI-enhanced condition significantly increased intrinsic motivation (p = .049), gameplay duration (p = .047), and frequency of NPC interactions (p < .001). These findings suggest that adaptive dialogic feedback may be more influential in shaping process-level outcomes, including motivation, sustained engagement, and exploratory interaction, than in producing immediate cognitive gains. The study contributes to educational technology research by positioning AI-driven NPC dialogue as a controlled mechanism for creating variable and socially responsive misinformation-evaluation practice in serious game environments.

Subscription-based educational technology (EdTech) companies experience significant revenue loss from student churn, making retention vital. Retaining engaged learners is generally more cost-effective than acquiring new ones. This study develops a predictive churn model using anonymised student data from an EdTech platform for children’s English learning. Following a CRISP-DM process, student activity, engagement, and financial features were engineered across multiple time windows to capture evolving student behaviours. Four machine learning algorithms-logistic regression, random forest, neural networks, and XGBoost (XGB)—were trained and compared. The optimised XGB model achieved the best performance, with approximately 0.84 accuracy, 0.63 F1 score, 0.85 recall, and the area under the curve (AUC) of 0.85 on test data, effectively identifying likely churners. Shapley Additive exPlanations (SHAP) based analysis revealed that engagement metrics, particularly the number of paid classes (both current and mean over 8 weeks), total learning time, engagement at gamified features, and student tenure, were the most influential predictors, confirming that highly engaged students are less likely to churn. This interpretable model provides actionable insights for retention strategies by predicting individual churn risk and highlighting key engagement drivers. In practice, even a 1% monthly reduction in churn could translate into multi-million-dollar annual savings for subscription EdTech providers. Overall, this research extends churn prediction into the EdTech domain, demonstrates the value of long-term engagement features, and applies explainable AI to enhance model transparency, thereby supporting its practical adoption.

Adaptive Artificial Intelligence (AI) platforms are increasingly promoted for personalised learning, yet their universal effectiveness is questionable, particularly as their design often overlooks the nuanced lived realities of atypical learners. This study addresses this gap by employing a participatory approach to engage students identified in the local school context as slow learners, not as test subjects, but as active co-researchers in evaluating ALEKS, an adaptive AI tutoring system for mathematics. Adopting a participatory usability framework within a rural Indonesian primary school, this qualitative case study centred on five students identified in the local school context as slow learners, positioning them as expert analysts of their own learning experiences. Through iterative think-aloud protocols and co-interpretive debriefing interviews, their navigation of the platform’s pedagogical design was collaboratively analysed. The co-analysis revealed three critical, learner-articulated phenomena: (1) a ‘survival-mode cycle,’ where punitive feedback mechanisms shifted the students’ focus from conceptual learning to system circumvention; (2) a ‘representational chasm,’ marking a cognitive and pedagogical failure in scaffolding the transition from concrete visuals to abstract symbols; and (3) a ‘linguistic disconnect,’ where formal system language created significant barriers to engagement and comprehension. By foregrounding learners’ voices, this study contributes a replicable, human-centred method for evaluating educational technology. The findings argue that for AI to be truly inclusive, its design and implementation must be fundamentally grounded in the participatory insights of its most vulnerable users.

This study investigates how motivational clusters (high, moderate, and low) influence self-efficacy, cognitive load, and emotional experiences within a metaverse project-supported ICAP (Interactive, Constructive, Active, and Passive) framework. In the highly immersive environment (i.e., the metaverse), both motivational factors and pedagogical approaches are essential to sustain students’ engagement. A total of 140 undergraduate students (80 males) in a cybersecurity course collaboratively designed a metaverse project using the Roblox platform, following the ICAP framework. Using statistical analysis and text-mining techniques, this study examined the interplay among motivation, students’ self-efficacy, cognitive load, and emotional responses. Results indicated that higher motivation was associated with greater self-efficacy, lower intrinsic cognitive load, and higher germane cognitive load, highlighting the role of motivation in shaping students’ confidence and mental load when dealing with complex tasks. Extraneous cognitive load was unaffected by motivational level, suggesting robust instructional design. Descriptive emotional analyses suggested that highly motivated showed relatively more frequent positive and negative emotional expressions, whereas students with lower motivation were more likely to report no emotional expressions. Further, students with low motivation showed slightly higher ratios for some positive emotions than moderately motivated students, suggesting that even students with lower motivation experienced enjoyment, curiosity, and trust in the novel technology. Topic modeling further suggested that positive emotional experiences were more prominent in the higher-motivation clusters, whereas negative emotions and cognitive challenges appeared more prominently in the low-motivation cluster. These findings underscore the critical role of motivation in regulating cognitive and affective processes in metaverse-based learning, providing both theoretical and practical insights for designing effective immersive educational environments.

This study examined how the use of gestures and embodiment influences the computational thinking (CT) skills and programming self-efficacy during unplugged debugging activities. A 2 × 2 factorial experimental design was employed, manipulating gesture type (congruent vs. incongruent) and embodiment type (direct vs. surrogate) during activities. Seventy-seven grade 2–3 students were randomly assigned and participated in debugging tasks, using either congruent gestures (aligning codes with movement) or incongruent gestures (arranging codes in a linear sequence), followed by direct embodiment (physically moving a character) or surrogate embodiment (observing a researcher move the character). Computational thinking was assessed using graphic-based measures of sequence comprehension, pattern recognition, and debugging, as well as transfer to text-based programming tasks. Data were analyzed using ANCOVA and ANOVA models, and findings revealed that incongruent gestures enhanced graphic-based CT proficiency, while surrogate embodiment improved the transfer to text-based programming, particularly in pattern recognition. Additionally, a significant interaction effect indicated that congruent gestures with direct embodiment produced the highest programming self-efficacy. These findings highlight the critical role of action-based learning in computational thinking, supporting unplugged activities as effective learning tools and reinforcing the need for inclusive curricula that enhance CT proficiency and programming self-efficacy in young learners. These findings suggest that different embodied debugging strategies uniquely support understanding, knowledge transfer, and efficacy, showing the importance of strategic integration of action-based activities in young learners.

The shift to online and blended learning, accelerated by the COVID-19 pandemic and the rise of generative AI tools, has intensified the search for valid and secure assessment methods. Oral assessments have long been recognized for their ability to authentically measure understanding, foster deeper engagement, and develop professional communication skills; however, logistical demands often limit their use. This paper reports findings from two studies investigating Asynchronous Oral Assessments (AOAs) delivered through a web-based platform and implemented alongside in-person, multiple-choice examinations. Study 1, conducted in an intermediate accounting course, explored associations between AOA participation and performance on multiple-choice exam items, revealing positive trends though not statistically significant. Study 2, conducted in a data analytics course, compared student performance across AOAs and in-person multiple-choice exams and found higher scores on AOAs, along with moderate correlations between formats. Survey results further indicated that students prepared differently for AOAs, reported greater use of active study strategies, and perceived the assessments as professionally relevant and cognitively engaging. Collectively, results suggest that AOAs represent an administratively scalable and pedagogically meaningful complement to traditional assessments.

Computer-supported collaborative argumentation (CSCA) creates valuable learning opportunities for students to explore multiple perspectives, co-construct knowledge, and develop argumentation skills. Grounded in collaborative learning and argumentation theories, we designed a CSCA activity to engage students in critical discussions of a controversial issue, with the goal of fostering perspective-taking, evidence-based reasoning, and critical evaluation of arguments. This activity included a series of structured tasks targeting key argumentation skills and gave student opportunities to explore both sides of the issue. We conducted a study with 64 middle school students to examine their dialogic patterns and argumentation skills in this activity. Results revealed a wide range of participation in the critical discussion across groups and individual students. Most collaborative efforts involved coordinating task responsibilities and processes among team members. Students’ argumentative discourse focused on developing and refining reasons and inviting teammates to provide ideas and elaboration. Interestingly, we found a negative correlation between the number of student dialogic turns and their scores on the collaborative argument task. This relationship can be explained by the differences across grade levels: older students performed better on the task despite engaging in fewer dialogic turns compared to younger students. By the end of their critical discussion, most groups proposed a team solution that integrated both sides of the issue. We conclude with implications for designing CSCA activities and support to promote middle school students’ argumentation skills.

This study analyzes linguistic patterns in reflective writing among graduate students in an instructional design course, using the Best Possible Self (BPS) method. The BPS exercise encourages students to envision their ideal professional selves, supporting early-stage professional identity development. We used Linguistic Inquiry and Word Count (LIWC) to examine psychological processes, emotional tone, social orientation, and temporal focus in students’ narratives. The analysis identified three reflective profiles—Causal Reflection-focused Individualists, Goal-Driven Pragmatists, and Social Collaborators—representing different approaches to identity development within a single course context. These profiles reflect emergent reflective patterns, not generalizable learner types. To assess whether professional identity development differed by cluster, we compared pre- and post-course survey responses. Results showed that Social Collaborators demonstrated significantly greater gains in professional identity scores compared to the other groups. While these findings are based on a short-term, context-specific study and do not imply causality, they suggest that socially oriented reflection may support identity growth. By integrating linguistic and narrative analysis, the study highlights diverse student approaches to constructing a professional identity. The results emphasize the value of adapting instructional strategies to support varied reflective orientations.

Although generative AI is increasingly integrated into K–12 education, prior research has emphasized post-intervention outcomes rather than how students interact with AI or how post-intervention competencies support human–AI collaboration. This mixed-methods study examined phase-based changes in students’ AI interactions, pre–post changes in AI dispositions, prompt engineering skills, and human–AI collaboration competencies, and predictors of post-intervention collaboration competencies. Sixty-nine eighth-grade students participated in a five-day STEM–AI curriculum using ChatGPT. Data from student-generated prompts, pre–post surveys, and competency tests were analyzed through content analysis, repeated-measures MANOVA, and multiple regression analyses. Results indicated that students’ AI interactions evolved from exploratory use toward argumentation and metacognitive monitoring. Students showed significant improvements in AI dispositions, prompt engineering skills, and human–AI collaboration competencies. Ethical awareness, particularly accountability and privacy, emerged as a significant predictor of post-intervention collaboration competencies. These findings suggest that generative AI can support higher-order thinking as a collaborative partner and that the development of human–AI collaboration competencies depends more on ethical awareness than on prompt engineering skills alone.

This study examines the use of ChatGPT as a design-support tool for developing career promotion content. In the digital transformation of society, areas such as career guidance and promotion are increasingly being reshaped by the speed, scalability, and accessibility of artificial intelligence (AI). Specifically, the study investigates how ChatGPT, a generative AI tool, can function as a design support system in the development of career promotion video content by generating structured interview scripts, profession-specific scenarios, and narrative components. The aim of the study is to identify the strengths and limitations of the ChatGPT-supported content development process, to evaluate the appropriateness of the content according to the feedback from professionals and students, and provide practical implications for educational content development and career guidance. Within the scope of the study, career promotion videos were developed for 38 different professions using three formats: interviews with professionals, a day in the profession scenarios, and critical decision moments. Data collection tools included researcher process notes, expert evaluation forms completed by professionals, and student feedback surveys following video implementation. Descriptive and thematic analyses were used to analyze the data. The findings indicate that ChatGPT-supported content development was perceived to enhance efficiency, scalability, and structured content generation, while human review and expert validation remained essential for ensuring contextual accuracy and professional authenticity. Overall, the findings suggest that conversational generative AI tools such as ChatGPT can serve as valuable design-support systems for educational video content development workflows, particularly in generating structured career narratives and scalable multimedia content. The study provides empirical and design-oriented insights into the integration of generative AI into educational content development and career promotion content development.

As programming education gains global importance, learners often struggle due to the cognitive demands, particularly in textual environments. Visual programming tools have been developed to ease this burden, but learners’ cognitive abilities—specifically executive function (EF)—may still affect their programming process and strategies. This study investigates how visual scaffolding influences the programming strategies of learners with different EF levels. Thirty-two adult participants were randomly divided into experimental and control groups, using programming systems with and without visual scaffolding respectively. Their EF was assessed using the Simon and running span tasks. Also, an eye-tracker was used to explore their cognitive processes during programming tasks and to identify their programming strategies. Results show that high-EF learners tended to adopt top-down strategies regardless of scaffolding. In contrast, low-EF learners without scaffolding used mixed strategies, while those with visual scaffolding were guided toward adopting a top-down approach. These findings highlight the role of visual scaffolding in supporting low-EF learners by fostering more effective programming strategies and mitigating cognitive differences between high- and low-EF learners.

School vacations disrupt learning continuity and often erode the momentum students need to stay engaged across time; more recently school closures due to the pandemic have produced similar outcomes. When students are not in school for an extended period, breaks in continuity and waning momentum often translate into irregular engagement with learning activities, allowing previously learned skills to grow rusty. The Keep in School Shape (KiSS) Program repurposes online survey software to provide students with convenient and engaging daily reminders and opportunities to maintain the math skills they need for their future studies via text message or email during times when school is not in session. This paper articulates the architectural logic of a low-stakes, voluntary review environment through a bounded design case and advances a set of transferable design commitments that can inform the design of similar interventions.

In this study, I examine The Great Zimbabwe (2021) board game as a pedagogical tool for teaching African archaeology and recent history through tabletop play. As digital and analog games increasingly shape 21st-century storytelling, their capacity to represent, reinterpret, and commodify the past demands critical analysis. Through analysis of the game’s design, mechanics, artwork, and player experience, I explore how The Great Zimbabwe constructs, represent, and transmit knowledge about ancient African civilizations. The findings indicate that the game embeds rich historical references within an economy-driven gameplay model. Players manage craftsmen, worship deities, trade cattle, and build monuments, mirroring the economic interdependence, religious pluralism, and political complexity that characterized ancient polities such as Great Zimbabwe, Mapungubwe, Mutapa, Kilwa, Lozi, and Zulu. Unlike most Eurocentric civilization board games that emphasize self-sufficiency, The Great Zimbabwe rewards cooperation, reflecting Africa’s historical trade networks and reciprocal political economies. Pedagogically, the game enables experiential learning by allowing players to simulate decision-making, economic trade, and spiritual life, transforming abstract history into lived experience. However, the study also highlights limitations inherent in Western-designed board games. The game compresses diverse African cultures, religions, languages, and histories into a homogenized fantasy, obscuring gender, kinship, and artistic nuances. Ultimately, the game occupies a productive but ambivalent space between history and imagination, serving both as a valuable educational tool and a reminder of the interpretive boundaries of play-based pedagogy.

This study explores the interactions between the quality of ICT access, ICT self-efficacy, math attitude, and math achievement utilizing the 2022 Programme for International Student Assessment (PISA) data for Hong Kong, Finland, and Türkiye. The study sample comprises 5190 participants from Hong Kong, 10,239 from Finland, and 7250 from Türkiye. The analyses based on structural equation modeling demonstrate that in all three countries, the quality of access to ICT significantly and positively affects both ICT self-efficacy and math attitude, which in turn affects math achievement. On the other hand, the direct effect of the quality of access to ICT is negative and significant on math achievement. Additionally, the economic, social, and cultural status (ESCS) plays a notable role in shaping ICT self-efficacy, math attitude (except Türkiye), and math achievement. ICT self-efficacy and math attitude mediate the relationship between ICT access and math achievement. The results underscore the importance of ICT access in enhancing students’ ICT self-efficacy and math attitudes and ultimately contributing to improved math achievement.

Ethics is recognized as a foundational element of human decision-making and a critical competency across educational disciplines, especially in STEM, where solving authentic, complex problems is essential. Traditional methods of ethics instruction often fail to provide the engagement needed for learners to grasp abstract ethical challenges. In contrast, game-based learning offers an experiential approach capable of supporting ethical literacy. Yet, despite their potential, games may also hinder learning or even promote unethical behavior, highlighting a critical but often overlooked challenge: balancing engaging gameplay with the meaningful integration of ethics-related objectives. Given the growing popularity and accessibility of analogue games in ethics education, this study explores their potential through four business ethics board games, each featuring distinct ethical goals, mechanics, and themes. A four-day workshop was conducted with 32 adult participants from diverse cultural, professional, and age backgrounds, with structured interviews following each gameplay session. By thematically analyzing participants’ experiences and perceptions, the study aims to deepen understanding of the complex interplay between games and ethics learning. Adopting a dialectical perspective, the research critically explores both the benefits and challenges of game-based approaches across different stages of the learning process. As a result, the pedagogical dilemmas related to rules, social modes, identity framing, systems nudging, personalized design, and ethics positioning were identified and discussed, spanning personal, behavioral, and environmental dimensions. Building on these insights, recommendations are proposed for designing and implementing games in ethics education, addressing aspects such as facilitation, instructions, goal settings, and social dynamics, thereby supporting meaningful, ethically informed learning experiences.