This study examined the impact of students' GenAI-use competence and agent interaction frequency on assignment performance, with perceived GenAI role and procrastination as mediators. We developed a statistical learning agent empowered by generative AI (GenAI) with Retrieval-Augmented Generation (RAG) to provide real-time interaction and valid instructional support for statistical learning tasks. A total of 35 students enrolled in an introductory statistics course participated. Interaction logs, along with surveys on students' perceived GenAI-use competence, perception of GenAI as a learning companion, and degree of learning procrastination, were analyzed. Results showed that higher interaction frequency with the agent was associated with better assignment performance. However, students with high competence in using GenAI tended to ask precise questions, resulting in fewer interactions with the agent. Additionally, procrastination negatively predicted interaction frequency, suggesting motivational barriers to engagement. The findings indicate that while GenAI agents can enhance student interactions and performance, individual differences in competence and procrastination significantly influence their usage. The study highlights the importance of aligning AI tools with students' cognitive and motivational needs. Future research should explore students' interaction patterns with AI agents to better facilitate their learning in statistics and analytics, optimizing the design and implementation of AI-assisted learning environments.

While digital learning offers advantages, it also presents challenges, including distractions from irrelevant websites. Such distractions characterize the personal learning environment (PLE), posing difficulties for learners and educators. Flow, an intrinsic motivation, is positively associated with learning outcomes, keeping learners engrossed and less influenced by external factors. Nevertheless, most prior research on flow has relied on surveys, and has overlooked the physiological aspect of flow during online learning. This study investigates the physiological signals of galvanic skin response (GSR) during flow experience in PLE. Using a natural online learning experiment, this study employed multimodal learning analytics to measure learners' cognitive processes objectively. The GSR, an indicator of cognitive load, was specifically used to measure and visualize physiological changes with different flow experiences in PLE. The results indicated that individuals with a high flow experience exhibited more stable emotional and stress responses than those with low flow tendency. This study is among the first to uncover the physiological shifts and stress reactions with flow experiences during the online learning process, offering a fresh perspective on flow theory via multimodal learning analytics.

To interrogate the inequitable amount of feedback provided by the instructor to each student in class, this study targeted at exploring the possibility of facilitating learning via peer feedback. Learning with peers is found to be beneficial to both disciplinary skills and learning motivation. Without time and space limits, educators can easily take advantage of social media develop online learning communities that empower students to conduct peer learning by giving feedback to each other. This research collected peer feedback in a Facebook private group of an introductory statistics course at a national university in northern Taiwan. Participants were 34 graduate students in the course. After analyzing and coding messages in the Facebook group into four levels (i.e., task, process, self-regulation, and self), this study found that students primarily received task-level feedback (40.79%) but infrequently received self-level feedback (8.48%). In addition, this study utilized machine learning techniques to examine the effect of different feedback levels on students' learning motivation. Results showed that self-regulation-level feedback stimulated autonomous regulation, but process-level feedback undermined it. From a student-centered perspective, this study proposed a practical framework promoting learning equity about receiving feedback. The present study implemented learning analytics, linking empirical evidence and motivational theory. It implies that teachers can promote learning equity by engaging students to initiate self-regulation level feedback for each other.

The use of computer-mediated communication (CMC) has been ubiquitous in higher education. To better understand students' behaviors and facilitate students' learning through CMC, this study aimed to classify messages in Facebook learning group which was created as an on-line discussion board. Different machine learning and deep learning classification models were proposed, trained and testified with corpuses from PTT, one of the famous on-line forums in Taiwan. Furthermore, the classification of Facebook messages by these well-trained models were compared with human coding. Results revealed that recurrent neural network (RNN) with word to vector (W2V) for feature extraction demonstrated the best performance in accuracy. In addition, the combination of RNN and TF-IDF was proved to have the highest correlation with human work. Implications for artificial intelligence (AI) in education context was discussed.

In the generation of educational reform, we have always wondering what kind of important ability has been taught in the curriculum? Financial literacy is an essential life skill (OECD, 2017). No matter what kinds of occupation will a person have, he or she must know how to manage his or her asset. In this research, we explored that if students’ math motivation, math anxiety, and financial concept would change or not after they received the hybrid financial courses in 4 weeks. We used math motivation and math anxiety as grouping variables, then use k-mean clustering to separate students into two groups. Two-way mixed-design ANOVAs were conducted to test the mean differences and change of math motivation, math anxiety, and financial concept across time points between different groups. We found only the significant change of math anxiety from students after they received the financial courses, while math motivation and financial concept kept constant.