The rapid integration of generative AI into education has raised important questions about its long-term influence on higher-order thinking. Building on our earlier study that examined the one-off impact of ChatGPT-4, the present follow-up study investigated its sustained effects on music students’ creative and analytical skills. A quasi-experimental design was employed with 74 undergraduate music students assigned to an experimental group (n = 36) and a control group (n = 38). Across a 24-week intervention, creativity and analytical skills were assessed at 12 repeated measurement points. Data were analyzed using latent growth modeling, multivariate analysis of covariance, cross-lagged panel modeling, epistemic network analysis, and a first-order Markov chain model. Results indicated that the experimental group began with higher initial levels and demonstrated steeper growth trajectories, reflecting a Matthew effect, while the control group displayed compensatory tendencies. Cross-lagged analyses revealed that analytical reasoning strongly predicted subsequent creativity in the experimental group, and network analysis confirmed stable co-occurrences between creative and analytical moves. Transition modeling further showed recursive cycles of idea generation, revision, evaluation, and structural reasoning. These findings suggest that ChatGPT-4 fosters an integrated interplay of creativity and analysis, offering music educators worldwide practical strategies for designing AI-mediated learning that balances innovation with critical rigor.
Children as young as three can engage in oral and visual planning when creating tangible objects from materials, yet it remains unclear whether such design competencies can be systematically strengthened in early childhood to cultivate future design thinkers and innovative creators. To evaluate a research-informed design thinking (DT) education program, we conducted a mixed-methods study. The quantitative component involved a quasi-experiment with 133 children (Mage = 5.22, SD = 0.86) from a Chinese kindergarten (intervention group: n = 68; business-as-usual comparison group: n = 65), where the intervention group received an eight-week DT program. Each child’s DT competence was assessed immediately before and after the intervention, followed by play-based interviews with 12 children (six per group). Child-level analyses indicated a statistically significant improvement in overall DT competence favoring the intervention group (Cohen’s d=0.38). However, gains were uneven across specific DT dimensions. Children showed more apparent gains in Empathize, Define, and Prototype, whereas no significant gains were observed in Ideate and Test. Furthermore, the youngest children appeared to show larger gains, whereas gender did not significantly moderate the intervention effects. Qualitative interviews provided insight into the cognitive processes underlying these developmental variations. Together, these preliminary findings may inform DT education research and practice in early childhood settings.
The ability to generate creative ideas is an important skill for education. This study examined how feelings of ease in generating ideas act as a cue that allowed individuals to have information to monitor their mental action and make important metacognitive decisions while generating creative ideas. Participants were asked to generate four ideas to help deal with the problem of empty work offices during and after the pandemic of COVID-19. After generating each idea, participants reported their feelings of ease and evaluated the creativity of each idea. Participants then reported their confidence in the ideas generated and selected their most creative idea. Results showed negative, curvilinear trajectories of feelings of ease and self-evaluations of the creativity of each idea. Feelings of ease in generating ideas were positively related to self-evaluations and the decisions to continue generating ideas and negatively related to choosing the first idea as the most creative. In addition, feelings of ease were indirectly and positively related to post-task confidence through their relationship with self-evaluations. These findings provided initial, empirical evidence that feelings of ease acted as cues to inform relevant judgments and decisions in the process of ideation.
Developing students’ mathematical creative thinking has become a central goal of mathematics curriculum reform worldwide. This study examines opportunities for mathematical creative thinking in the three Grade 1 Chinese mathematics textbook editions first implemented under the Compulsory Education Mathematics Curriculum Standards (2022 Edition). Drawing on the Opportunities-to-Learn framework, the study localizes Hadar and Tirosh’s (2019) analytical framework by refining the operational indicators of divergent, convergent, and lateral thinking and extending the framework through the incorporation of cognitive complexity. A human–AI collaborative coding approach was used to analyze 1,473 mathematical tasks. The findings indicate that approximately half of all textbook tasks provide opportunities for mathematical creative thinking, with convergent thinking emerging as the dominant dimension across all three editions, although each edition places emphasis on different indicators of mathematical creative thinking. This study contributes a localized analytical framework for evaluating mathematical creative thinking in curriculum materials and provides implications for textbook design, curriculum reform, and future research on mathematical creative thinking.
Developing systems thinking through coordinated reasoning across multiple abstraction levels remains a persistent challenge in microelectronic circuit design education. Students often struggle to connect high-level system specifications with subsystem interactions and low-level transistor implementations, particularly when addressing complex integration and debugging tasks. While simulation tools are commonly used to support circuit analysis, their role in bridging abstraction levels within project-based learning contexts remains underexplored.This study examines how simulation-based project learning, implemented through LTspice, supports students’ ability to navigate abstraction levels and develop systems thinking in a third-year microelectronic circuit design course. Over a 12-week semester, 49 undergraduate electrical engineering students engaged in a sequence of scaffolded design projects culminating in the construction and integration of an analog-to-digital converter. The projects required iterative movement between system-level architecture, functional block design, and transistor-level implementation, supported by continuous simulation-based testing and refinement.A qualitative–performance research design was employed, drawing on students’ project artifacts, design reports, final practical examinations, and instructor observations. The analysis reveals a shift from isolated component-focused reasoning toward system-level debugging and cross-level explanations of circuit behavior across successive project phases. Later design reports frequently articulated how lower-level design decisions influenced overall system performance, particularly during integration and optimization phases. Performance outcomes further suggest improved effectiveness in diagnosing system-level errors and coordinating design trade-offs across abstraction layers.The findings suggest that simulation-based project learning can function as a mediating mechanism that supports coordination between abstraction levels during system integration tasks. This study contributes to engineering education research by framing simulation as a mediator for systems thinking development in microelectronic circuit design and by offering instructional design principles for structuring project-based learning environments that explicitly target abstraction-level transitions.
The integration of computational thinking (CT) into science, technology, engineering, and mathematics (STEM) education is often approached through external interventions, including curricula, technological tools, and policy initiatives. This study examined how teachers interpret these contextual conditions and translate their understanding of CT into instructional practice. Using a Straussian grounded theory approach, one-to-one semi-structured interviews were conducted with 13 middle-school science, mathematics, and computer science (CS) teachers in Türkiye. Data were analyzed iteratively through open, axial, and selective coding, as well as constant comparison. The analysis generated the Exploring the Generative Gateway to CT (EGG) Model, a context-specific substantive framework comprising three interconnected domains: external factors, internal factors, and learning-design factors. External factors included instructional, managerial, organizational, social, and technological contexts. Internal factors concerned teachers’ awareness, pedagogical interpretations, emotions, commitment to self-development, and risk of professional stagnation. Learning-design factors comprised the competencies and professional development processes through which teachers designed, adapted, implemented, and assessed CT-integrated STEM activities. The findings indicate that external conditions enable, constrain, or redirect teachers’ practice, while teachers’ interpretations and learning-design competence shape how they respond to those conditions. Teachers sometimes adapted activities, sought resources, collaborated with colleagues, or pursued professional learning; however, these actions mitigated particular difficulties rather than eliminating structural barriers. Sustainable CT integration, therefore, requires complementary support for teachers’ conceptual understanding and learning-design competence together with adequate curricular, institutional, collaborative, and technological conditions.
Introduction Fostering creativity and self-esteem in childhood are very important factors in child development, and the relationships between the two are established throughout the different stages of development. The objectives of the research were (a) to analyse creativity in children aged 3 to 6; (b) to understand how gender, geographical context, school year, age and participation or non-participation in extracurricular activities influence the development of creativity; and (c) to determine whether there is a relationship between levels of self-esteem and creative ability. Methodology This study analyses creativity and its relationship with self-esteem in 137 children in early childhood education. A sample from two schools (urban and rural) in Granada, Spain, was evaluated using the Creative Thinking Test and the Questionnaire for the Assessment of Self-Esteem in Childhood. After verifying the normality of the data, inferential analyses were performed (Mann-Whitney U test and Kruskal-Wallis test), calculating the effect size. Results The results showed that girls and rural students scored higher in creativity. Furthermore, creativity increased with age, school year and participation in extracurricular activities. Correlations have been established between creativity and self-esteem, especially in the academic and general dimensions. Conclusions The study concludes that fostering creativity and self-esteem from an early age is key to children's emotional, cognitive, and social development. As practical implications of the study, the importance of fostering development by providing a safe and motivating environment, both at school and at home, was highlighted. Limitations of the study included sample size and the specific context of the study.
Family and school contexts are central to adolescent development, yet their joint and nonlinear contributions to creative self-efficacy remain underexamined. Based on the Ecological Systems Model of Creativity Development, this study investigates how the creative family environment and the creative school environment jointly predict adolescents’ creative self-efficacy, with growth mindset as a potential mediator. We analyzed data from 10,291 secondary students in Hong Kong and Macao from PISA 2022 (48.3% female; Mage = 15.73 ± 0.44) using response surface analysis and structural equation modeling. Three findings stand out. First, both family and school contexts positively predict creative self-efficacy, and when support is high in both contexts the increase follows an accelerating nonlinear pattern, consistent with positive curvature along the line of congruence. Second, under contextual incongruence, the school context shows a compensatory effect and can offset weak family support. Third, growth mindset partially mediates the effects of family and school contexts on creative self-efficacy. No significant gender differences emerged in either mean levels or structural pathways. These results highlight the importance of coordinated efforts between families and schools, identify growth mindset as a central psychological mechanism, and offer actionable guidance for building aligned ecosystems that reliably strengthen adolescents’ creative self-efficacy.
Artifact-based assessment assumes that the quality of a student's work reflects the student's underlying expertise. Large language models (LLMs) break this assumption by decoupling output quality from user knowledge, which renders artifact-only evaluation diagnostically unreliable. This shift also exposes a long-standing imbalance in engineering education, in which the profession has relied on both production and review, yet curricula and grading have emphasized production far more than the skills that make review effective. In engineering, review is the disciplined detection and correction of errors that can propagate into unsafe designs, incorrect calculations, or noncompliant decisions. Historically, teaching review at scale has been constrained by the scarcity of high-quality, realistic, and flawed artifacts. LLMs invert this constraint by generating plausible‑but‑wrong engineering work on demand, at controllable levels of subtlety and difficulty. Consequently, we argue that assessment validity can be strengthened by treating review performance in terms of error detection, diagnosis, and justification as a primary learning outcome rather than treating artifact production as the primary proxy for competence. We position review competence as a domain-specific form of critical thinking and evaluative judgment, anchored in the constraint structures that distinguish engineering verification from generic critique. Building on this perspective, we present a taxonomy of LLM-generated engineering errors, categorized by error type, severity, and detectability. We also provide templates for prompts and an accompanying LLM interface for generating on-demand, review-centered instructional activities.
Creativity in addressing social issues requires the integration of diverse perspectives for sustainable and ethical solutions, a concept known as Transformational Creativity. To explore how such imagination can be fostered, this study focuses on action-based imagination. In this experiment, we compare two conditions: one involving the physical imitation of others and another involving the recall of existing knowledge about social issues. A 2 × 2 between-participants factorial design is employed, manipulating the activation of Physical Imitation (present/absent) and Memory Recall (present/absent), with 80 participants randomly assigned to one of 4 conditions. Dependent variables are empathy and a sense of engagement with social issues, measured before and after the intervention. The results show that empathy significantly increased from pre- to posttest in conditions involving physical imitation, indicating its effect. In contrast, no significant differences were found in social issue engagement across conditions. These findings suggest that action-based imagination provides a foundation for cultivating empathy in the context of social issues, which is a foundational process underlying Transformational Creativity. They also offer practical insights for designing educational practices that prepare citizens to contribute to a sustainable and diverse future.
In foreign language writing instruction, model texts are frequently treated as normative templates for replication rather than as resources for creative exploration. Creative imitation addresses this tension by guiding learners to analyze and transform model texts to generate original expression, yet research on scaffolding this process remains limited and concentrated in English-language contexts. This study integrates SCAMPER as cognitive scaffolding for creative imitation through a 16-week mixed methods classroom intervention in a Japanese writing course at a Chinese university. The experimental group (n= 28) received creative imitation instruction with SCAMPER scaffolding, while the comparison group (n= 29) received form-focused instruction. The experimental group significantly outperformed the comparison group in content and organization, in flexibility, originality, and elaboration of creativity in writing, and in domain-general divergent thinking, although gains in convergent thinking were limited. Longitudinal interviews revealed gradual cognitive, linguistic, and affective development, with gains in self-efficacy and intrinsic motivation particularly salient among lower-proficiency learners. These findings suggest that imitation, with appropriate cognitive scaffolding, can serve as a productive foundation for creative generation, and offer empirical support for the creative imitation framework in foreign language contexts beyond English.
Although graduate students frequently encounter illegitimate tasks—assignments perceived as inconsistent with their research roles—little is known about how such tasks influence research creativity and through what cognitive mechanisms. Drawing on Cognitive Appraisal Theory (CAT), a dual-pathway model is proposed in which illegitimate tasks influence research creativity through problem-solving pondering and affective rumination, with supervisor–student relationship quality as a moderator. Data were collected from 389 graduate students in Chinese universities using a two-wave survey design, and the hypothesized moderated mediation model was tested using hierarchical regression analyses supplemented by the PROCESS macro. The results indicate that illegitimate tasks exert opposing indirect effects on research creativity through distinct reflective cognitive pathways. Illegitimate tasks were associated with increased affective rumination, which interfered with sustained creative engagement. However, under high-quality supervisor-student relationships, illegitimate tasks were more strongly related to problem-solving pondering, which facilitated creative thinking. Supervisor-student relationship quality further shaped how illegitimate tasks were interpreted and which reflective pathway was activated. These findings clarify the cognitive mechanisms and relational conditions through which illegitimate tasks influence research creativity and contribute to understanding how task legitimacy shapes creative thinking in graduate research contexts.
This study utilized survey data from 1,257 Chinese eighth-grade students to examine the associations among teaching approaches (i.e., inquiry-based teaching and teacher-directed instruction), epistemological beliefs, utility value, and science-related career choice. Crucially, these relational patterns were analyzed within different levels of metacognitive ability (high, medium, and low). Specifically, the present study employed multi-group structural equation modeling to examine group-specific patterns of associations among variables across students with different levels of metacognitive ability. Results indicated that positive associations between inquiry-based teaching and the examined variables were mainly observed within the high-metacognitive group, whereas teacher-directed instruction demonstrated positive associations with certain variables across groups. In addition, indirect associations involving utility value in the relationship between teaching approaches and science-related career choice were identified in the high-metacognitive group, while comparable patterns were not observed in the other groups. These findings suggest that the association patterns involving teaching approaches, epistemological beliefs, utility value, and science-related career choice may differ across metacognitive groups.
Innovation often comes from posing unconventional problems, while disciplinary integration requires formulating interdisciplinary problems. Although studies have investigated the association between cognitive style and general academic achievement, the influence of cognitive determinants on domain-specific problem-posing competence (PPC) remains underexplored. This study used the Embedded Figures Test to categorize students as having either a field-dependent (FD) or a field-independent (FI) cognitive style. Over a 16-week course, we collected generative text-based problems from 90 high school students aged 15–17 years (M = 15.94, SD = 0.63) across eight learning topics. Epistemic network analysis (ENA) was conducted to explore the characteristics and development trajectories of PPC. Results showed that the problems posed by FD students exhibited stronger connections between fluency and elaboration, typically targeting factual knowledge at the cognitive level. By contrast, the problems posed by FI students exhibited greater associations with flexibility and originality, including more open-ended and integrative problems. Furthermore, as the course progressed, FD students benefited more from the learning materials, which led to significant improvements in their cognitive level and the originality of their problems compared to FI counterparts. These findings indicate that PPC is a dynamically evolving competence throughout the learning process. Educators should design adaptive instructional scaffolds tailored to cognitive styles, and encourage learners to engage in continuous metacognitive reflection on their thinking habits and problem-posing processes.
This study examines the perceived influence of Generative Artificial Intelligence (GenAI) on Higher Order Thinking Skills (HOTS) in higher education through a multi-actor analysis at Universidad de los Andes, Colombia. Using a concurrent mixed-methods approach, we conducted nine semi-structured interviews with academic administrators, twelve focus groups with sixty-five faculty members, and surveyed 529 students. Qualitative data were coded for four HOTS: critical thinking, creative thinking, reasoning, and metacognition. Findings reveal that concerns about creative thinking were the most frequently mentioned (31%), followed by general cognitive skills (25%) and critical thinking (24%), while reasoning (11%) and metacognition (9%) received less attention. Notably, patterns of concern varied significantly across stakeholder groups: students expressed greatest worry about creative thinking (50% of coded entries), fearing loss of creative capacity through AI dependency; faculty showed more balanced concerns and openness to using GenAI as a creativity catalyst; academic administrators prioritized critical thinking (30%) and reasoning (26%), linking these to digital literacy and institutional quality. Results suggest the need for developmental frameworks that consider when and how AI assistance is pedagogically appropriate, multi-level institutional responses addressing divergent stakeholder concerns, and ongoing dialogue about balancing efficiency gains with cognitive development. The research contributes to understanding GenAI’s complex, sometimes contradictory impacts on thinking skills in higher education.
Addressing Sustainable Development Goals (SDGs) has become a pressing global responsibility. Critical-thinking dispositions (CTD) play a pivotal role in fostering creative problem-solving (CPS), particularly when tackling complex social challenges. To promote meaningful learning in this context, the present study designed an integrative mindfulness intervention—Listen Mindfully, Watch Mindfully, and Observe Mindfully—integrated with SDG-based critical thinking activities to enhance college students’ CTD and CPS abilities. This study also examined the relationships among mindfulness, CTD, and CPS ability. To achieve our goals, we developed two instruments: the Inventory of Critical-Thinking Dispositions (ICTD) and the Inventory of Creative Problem-Solving Ability (ICPSA). Reliability and validity analyses were conducted using data from 434 college students for the ICTD and 460 college students for the ICPSA. For the intervention design, 82 college students were assigned to either an experimental group (which received the mindfulness intervention) or a control group. Results indicated strong reliability and validity for the developed instruments. Moreover, the mindfulness intervention significantly improved the experimental group’s CTD and CPS ability. Process model analysis further revealed that CTD mediated the effects of mindfulness on CPS ability in the context of SDG-related problem-solving. This study introduces an innovative mindfulness approach—integrating meditative and creative mindfulness—to enhance CT dispositions and CPS. The findings further emphasize the importance of incorporating SDG issues into CT practices to foster essential 21st-century competencies.
Generative AI is increasingly used in academic writing, yet little guidance exists on how to integrate it without replacing the author’s reasoning or diluting their scholarly voice. Accordingly, the paper is positioned as a design-oriented methodological framework that functions as a metacognitive scaffold by specifying an operational procedure for post-draft AI use rather than as an empirical study or a purely theoretical contribution. In this proposed framework, authors first produce a human-written draft and only then engage generative AI as a diagnostic conversational companion restricted to post-draft auditing of coherence, alignment, and reconstructability. In this revision stage, generative AI is used to perform specific, verifiable actions: checking coherence across sections, auditing alignment between research questions, methods, results, and conclusions, detecting ambiguous or overly abstract formulations, and verifying adherence to reporting conventions. Drawing on Maturana’s concept of structural coupling, we argue that this boundary preserves the author’s epistemic authority by ensuring that the meaning of the manuscript originates in the author’s own distinctions, while AI contributes through perturbations that prompt clarity and refinement rather than generating content. The framework expands current guidelines by moving beyond general ethical prescriptions to define a practical, auditable procedure for post-draft AI use that preserves authorial voice and epistemic control. It offers a replicable workflow that integrates efficiency with transparency, helping researchers use generative AI to strengthen—not substitute—their scholarly reasoning, transforming the role of AI from a content generator into a catalyst for critical thinking and creative precision.
Despite the efficacy of peer feedback in enhancing students’ writing performance, its application is hindered by unstable feedback quality. Artificial intelligence (AI) offers new possibilities for this feedback mode, yet most research has largely focused on comparing the effectiveness of these two feedback types rather than examining their integrated application. Grounded in Activity Theory, this exploratory case study investigates the effects of peer-mediated AI-generated feedback on continuation writing among Chinese senior high school students. Quantitative results reveal this hybrid feedback model significantly improves students’ writing accuracy, while no statistically significant gains are observed in complexity and fluency. Qualitative analysis further demonstrate a dynamic feedback processing process: students critically triangulate AI suggestions with peer insights and personal judgment, which effectively redistributes the workload of feedback provision. The study also identifies key contradictions within the activity system, especially the tension between the innovative practice and the exam-oriented culture in China. We argue that AI should be reconceptualized as a mediating tool, which enables peer feedback to develop into a more competent collaborative pedagogical practice.
This study reports the development and initial validation of a relational reasoning test for middle school students in science/STEM education. The instrument consists of diagnostically branched tree items, scored 0–3, representing the first empirically evaluated implementation of this format in the assessment of relational reasoning in science contexts. The Relational Reasoning Skills Test (RRST) is grounded in Piaget’s theory of formal operations. Evidence for its psychometric properties was examined using classical test theory analyses, including item difficulty and discrimination indices. The final version consists of 10 items, with a Cronbach’s alpha coefficient of .70, which is generally considered acceptable for early-stage instrument development. Exploratory factor analysis indicated a two-factor solution explaining 44.60% of the variance, with factor loadings ranging above .40. Confirmatory factor analysis results indicated an acceptable model fit for the proposed structure. Overall, the findings suggest that the RRST may provide preliminary evidence for assessing relational reasoning in science contexts. However, further validation studies with larger and more diverse samples are required to strengthen the psychometric support and generalizability of the instrument. Future research may also explore the integration of diagnostically branched items into adaptive and technology-enhanced assessment systems.
