Gonzalo Lorenzo

Department of Development Psychology and Teaching, University of Alicante, Spain // glledo@ua.es

Alejandro Lorenzo-Lledó

Department of Didactics and School Organisation, University of Granada, Spain // alorenzolledo@ugr.es 

Isabel Góméz-Barreto

Department of Pedagogy, University of Castilla la Mancha, Spain // isabelmaria.gomez@uclm.es

María Teresa Bejarano-Franco

Department of Pedagogy, University of Castilla la Mancha, Spain // mariateresa.bejarano@uclm.es

Andrea Cerdán-Chacón

Department of Development Psychology and Teaching, University of Alicante, Spain // andrea.cerdan@ua.es 

ABSTRACT: 

In recent years, robotics has emerged as a powerful and appealing instrument for autistic children, who struggle with communication. This study aims to analyze the improvement in social communication and interaction derived from the intervention with the NAO robot. A quantitative methodological approach has been adopted together with a quasi-experimental design based on a paired pre-test–post-test model. The participants were sixteen autistic children at the curriculum competence level of kindergarten (three-year-olds) and second grade and classified as ASD levels 1 and 3. The intervention consisted of eleven sessions, in which activities such as identifying the cause of a mood in a social context were proposed. The main instruments used for data collection were the modified checklist from the Early Start Denver Model (ESDM) and a field journal. The results showed substantial improvements in expressive communication, joint attention behavior, and social skills during interactions with both adults and peers. Cohen’s d values, close to or greater than 0.8, support these differences, indicating that the improvements observed are both visible and meaningful. Therefore, it can be concluded that the NAO robot may be a valuable tool for developing expressive communication, joint attention behavior, and social skills in interactions with both adults and peers. Consequently, it is recommended to progressively incorporate it into school environments to address these dimensions where autistic children require support. For future studies, it would be advisable to increase the sample size to potentially achieve more substantial improvements in other dimensions where no significant improvements were found.

Keywords: 

Robotics, NAO, Autism, Communication, Social interaction  

Xin Gong

College of Education, Capital Normal University, China // Gongxinjyjs@163.com

Shufan Yu

Department of Chinese and Bilingual Studies, Faculty of Humanities, The Hong Kong Polytechnic University, Hong Kong, China // yushufan1993@gmail.com

Xiu Guan

Institute of Education, Tsinghua University, China // guanx23@mails.tsinghua.edu.cn

Ailing Qiao

College of Education, Capital Normal University, China // Qiaoal@126.com

ABSTRACT: 

Prior studies have mainly focused on testing collaborative programming learning (CPL) patterns while neglecting the exploration of the dynamic evolution of social epistemic interaction patterns among different groups. Studying the social and epistemic network nature of learner interaction is crucial to understanding the CPL process. This study aims to explore the social epistemic interaction patterns and their evolutionary path among different groups. In this quasi-experimental design, 51 high school students were randomly allocated into 17 groups. Content analysis was used to analyze online collaborative conversations and interaction contents in the early, middle, and later periods of CPL. Social epistemic network and cluster analyses revealed three interaction patterns. The results showed that groups in cluster 1 were composed of core roles, which exhibited a multi-center balanced collaboration pattern (MBCP), and their social epistemic interaction levels showed a continuous upward trend; groups in cluster 2 included core, semi-core, and edge roles, respectively, and demonstrated a hierarchical center-led coordination pattern (HLCP) that initially gained but later declined in social epistemic interaction levels; groups in cluster 3 included core and edge roles, and displayed a single-center feedback cooperation pattern (SFCP), which remained consistently low in social epistemic interaction levels. Our findings emphasize the importance of CPL’s social epistemic interactions. By recognizing these patterns, educators can better facilitate meaningful student interactions, fostering deeper learning and social development.

Keywords: 

Collaborative programming learning, Social epistemic network analyses, Interaction patterns, Evolutionary path

Yicheng Sun

School of Mechanical and Precision Instrument Engineering, Xi’an University of Technology, China // sunyicheng@stu.xaut.edu.cn

Yi Wang

School of Mechanical and Precision Instrument Engineering, Xi’an University of Technology, China // School of Art and Design, Xi’an University of Technology, China // wy2005@xaut.edu.cn

Hanbo Yang

School of Mechanical and Precision Instrument Engineering, Xi’an University of Technology, China // yanghanbo@stu.xaut.edu.cn

Richard Suen

Faculty of Management, Shenzhen MSU-BIT University, China // 2120240147@smbu.edu.cn

ABSTRACT: 

With increasing student enrollment in higher education, timely and effective analysis of student feedback has become essential for course evaluation and instructional improvement. Traditional methods often struggle with complex feedback containing mixed sentiments and lack the adaptability to accommodate evolving feedback entries. To address these limitations, we propose a novel framework that integrates Large Language Models (LLMs) with an adaptive template-based cache for comprehensive feedback analysis. By leveraging the semantic understanding capabilities of LLMs and a dynamic caching mechanism, our approach continuously updates and refines templates in response to new feedback themes. It employs hierarchical matching for fine-grained classification and uses In-Context Learning (ICL) to identify semantically relevant templates, thereby enhancing summary generation while minimizing redundancy. The Feedback Summary Generation component synthesizes sentiment trends and category-level distributions into actionable reports to support data-driven decisions in educational settings. Empirical results demonstrate the superiority of our framework in terms of summary quality, information coverage, sentiment classification, and processing efficiency, offering a robust and adaptable solution for feedback analysis across diverse educational environments.

Keywords: 

Student feedback, Large language model, Adaptive template-based caching, In-context learning, Cache matching, Cache updating

Ashraf Sadat Ahadzadeh

School of Communication, Xiamen University Malaysia, Malaysia // ashrafsadat.ahadzadeh@xmu.edu.my

Shin Ling Wu

School of Psychology, Faculty of Medical and Life Sciences, Sunway University, Malaysia // shinling_wu@hotmail.com

Changsong Wang

School of Communication, Xiamen University Malaysia, Malaysia // cswang@xmu.edu.my

ABSTRACT: 

ChatGPT is the most cutting-edge AI language model with the potential to transform education. Building on the Theory of Planned Behaviour (TPB), this study aims to investigate how attitude, subjective norms, perceived behavioural control influence the intention to use ChatGPT for academic purposes, aiming to expand upon the existing literature. Furthermore, drawing upon the Cognitive Moral Development Theory, it explores the moderating effect of ethical judgments on the relationship between TPB constructs and the intention to use ChatGPT for academic purposes. A cross-sectional survey was utilised to gather 311 responses from university students studying in Malaysia. We discovered that the intention to use ChatGPT was positively influenced by TPB constructs while negatively affected by ethical judgments that discourage the use of ChatGPT for academic purposes. Furthermore, these ethical judgments did not attenuate the association between attitude and intention. Individuals with higher scores on ethical judgments disfavouring ChatGPT exhibit a weaker link between perceived behavioural control and their intention to use this technology. Ethical criticisms of ChatGPT usage strengthens the link between subjective norms and intention. These findings provide practical implications for incorporating generative AI tools in education, underscoring the ethical utilisation of these technologies, aiming to leverage their advantages while minimising their potential drawbacks. The lack of established causal links between variables in the study suggests the use of experimental designs in future research. Moreover, investigating ethics as a complex concept and fusing ideas from psychology and education may provide a deeper understanding of ChatGPT adoption.

Keywords: 

ChatGPT, Ethical judgments, Attitude, Subjective norms, Perceived behavioural control