Educational robotics (ER) has great potential to develop both computational thinking (CT) and 21st-century skills in higher education (HE). However, structured frameworks that help educators align platform selection with desired learning outcomes are still lacking. This study presents a competency-oriented framework that organizes ER platforms along a functional continuum based on the dominant mode of student interaction. Rather than applying rigid classifications, the framework places platforms on a spectrum ranging from screen-based programming environments to systems in which behavioral logic is integrated into the physical morphology of the robot. Based on the analysis of thirty-two empirical studies, the framework connects interaction modalities with groups of CT dimensions and cross-cutting skills, such as creativity, collaboration, adaptability, leadership, and ethical reasoning. CT is conceptualized as a scaffold that mediates the development of these skills through embedded interaction with robotic systems. The resulting structure offers a practical tool for curriculum planning, platform selection, and instructional design in university settings. By mapping how different types of ER platforms support specific learning processes, the framework contributes to a more intentional and evidence-based integration of robotics in HE, aligned with educational goals and workforce demands across disciplines.
Generative artificial intelligence (GenAI) tools have introduced promising opportunities for enhancing learning and automated assessment. Self-directed learning (SDL) and a design thinking mindset (DTM) have also been recognized as promising pedagogical approaches that support and contribute to advancing the Sustainable Development Goals. Although prior research has suggested that components of SDL may influence motivation and that motivation may be associated with DTM ability, there remains a noticeable gap in empirical studies examining the interrelationships among these variables. Grounded in the principles of SDL, the present study investigates the use of GenAI tools in assessing students’ DTM. A quantitative study was conducted with a sample of 379 university students, and the data collected through validated measurement instruments were analyzed using structural equation modeling to test and verify the proposed structural model. Empirical findings revealed that the use of GenAI tools for assessment significantly influenced all components of SDL, ultimately leading to enhanced learning motivation. Furthermore, awareness, learning strategies, learning activities, and interpersonal skills were found to mediate the relationship between the use of GenAI tools in assessment and students’ learning motivation. Interestingly, students’ self-assessment did not significantly affect learning motivation and did not play a mediating role between GenAI tools-based assessment and motivation. A particularly noteworthy finding is the absence of a significant relationship between motivation and DTM. This study highlights the role of GenAI tools in relation to students’ learning motivation and the development of their DTM, offering important implications for policies governing virtual learning environments and for instructional practices.
Studying the cognitive engagement shown in forum messages of informal online communities can help to associate discussion behaviour with underlying learning processes. The ICAP framework can serve as a theoretical lens for labelling forum messages by distinguishing different levels of cognitive engagement, yet requires time-consuming, high-inference human coding, which is not feasible at scale. Although this can be mitigated by supervised classification models, these rely on large amounts of labelled training data, and their decision-making often remains opaque. In contrast to this, Large Language Models (LLMs) work “out-of-the-box” for qualitative coding tasks via natural language instruction and can produce reasoning traces and explanations for their judgments. In this study, we explore this potential by integrating the ICAP framework into natural-language prompts for few-shot in-context classification, comparing LLM judgments and explanations with those of human raters. Results indicate that LLM judgments generally align with the human ratings, while their reasoning traces and variations offer additional insight into where the coding scheme invites ambiguous interpretations. This suggests that LLMs could serve as cognitive proxies that help diagnose and improve high-inference coding processes. Finally, we apply our approach to large-scale forum data, revealing cross-platform differences in expressed cognitive engagement.
Predictive methods play a key role in learning systems to enable effective interventions. However, most existing approaches emphasize concept-level suggestions and pay limited attention to cognitive constructs aligned with instructional goals. For instance, in mathematics, mastery depends on the integrated and balanced development of multiple interwoven proficiency strands. When predictive outputs do not reflect instructional objectives, educators may struggle to trust or utilize the insights generated by these models. To address this, we propose a predictive framework that incorporates cognitive construct-aligned features into the modeling process. The research focuses on elementary mathematics, where proficiency strands represent fundamental cognitive dimensions of learning. Since manually labeling these strands is time-intensive for experts, large language models (LLMs) are used to classify semantic features. These features are integrated into the proposed Proficiency-informed Knowledge Tracing (PKT) model as auxiliary signals to enhance mastery estimation. PKT is evaluated on the XES3G5M dataset, which contains elementary mathematics problems, and is compared with established deep learning models. The results indicate that PKT achieves overall competitive performance and surpasses base models, demonstrating a capacity to capture latent mastery patterns. Further experiments reveal that integrating LLM-generated features improves model calibration, as measured by Mean Absolute Error (MAE) and Root Mean Square Error (RMSE), compared to base models. PKT ranked first overall when predictive performance and calibration were considered together, followed by AKT and sparseKT. Additionally, the model generates interpretable diagnostic profiles at both cognitive and concept levels. These findings underscore the value of incorporating cognitively relevant constructs into knowledge tracing.
In formative assessment (FA), students’ agency is increasingly emphasized, making it essential to provide feedback that is both actionable and reflective of their learning processes. However, existing learning analytics dashboards (LADs) often fall short in delivering such feedback. To address this gap, we proposed the concept of “social process awareness” (SPA), which encourages students to reflect on the learning process of their peers. This study explores the feasibility of implementing SPA by analyzing weekly test scores and unit exam performance of grade 8 students in a Japanese junior high school. Through clustering analysis, we identified four distinct learning process patterns and examined their distribution across performance groups. The results suggest that students could benefit from referencing the learning processes of peers who perform slightly better, enabling stepwise improvement. Additionally, the study reveals that common learning challenges can be identified across all groups, suggesting the value of visualizing shared difficulties. These findings demonstrate the feasibility of SPA and provide insights for designing LADs that support student-led FA in K-12 education.
