Profiling readers provides valuable insights into the interaction between affective and cognitive engagement in multiple-text reading. However, metacognition-particularly goal-oriented and problem-awareness metacognition-remains underexplored. This study integrates unsupervised machine learning (k-means clustering) and structural equation modeling (SEM) to examine how these metacognitive factors interact with affective engagement (topic interest and self-efficacy) and their relationship with reading processes and writing performance. A total of 152 participants read four texts on genetically modified technology and wrote a synthesis report. Their reading behaviors were recorded using log data and eye-tracking technology to capture the integration process, while writing performance was assessed through multiple-text integration and source citation. Results revealed three distinct reader profiles: high-metacognition, high-interest, and disengaged groups. SEM analysis showed that, compared to the high-interest group, the high-metacognition group exhibited greater integration behaviors, which in turn enhanced multiple-text integration and source citation. Moreover, the integration process successfully mediated the relationship between reader profiles (High-metacognition vs. High-interest) and writing performance. These findings highlight the critical role of goal-oriented metacognition in multiple-text reading and emphasize the need to further explore metacognitive factors in comprehension research.

Emotion plays an important role in online reading and self-regulated learning. It could influence reading processing and even the outcomes through motivation functions. Since previous studies mostly adopted questionnaires to access readers’ emotions, research that applied facial emotion detection non-intrusively during reading process is still limited. Therefore, the purpose of this study was to investigate how readers’ on-going emotions fluctuated under different reading conditions, i.e., with different reader characteristics and reading orders, using a facial emotion analysis system. 21 participants were recruited and presented with four conflicting scientific texts that included positive or negative arguments on genome editing. These participants were divided into the high and low engagements and then were assigned different reading orders of positive-first or negative-first articles. The results showed that participants who read negative-first expressed more sadness and less happiness. Meanwhile, the high engagements showed more anger when reading positive-first, but the low-engagements showed more anger when reading the negative-first. Reading negative-first performed better, but writing performance did not differ according to the group membership or reading order. These findings suggest that (1) the influence of emotions on online reading multiple and conflict science texts and performance is complex; negative activating emotions like sadness and anger could be beneficial to reading performance, while positive activating emotions like happiness could be harmful for reproductive performance. (2) Using facial expression to non-intrusively detect readers’ momentary emotions provides meaningful insights into the fluctuation of emotion during reading science texts. With only six basic emotions, it may not be enough to fully show the motivational transitions in reading process. The interpretation of these facial emotions should be considered beyond their surface meaning.

To assess junior high school students' engagement in collaborative STEM project-based learning courses, this study employed a web-based observation protocol called COPIE-STEM to collect classroom observation data. This protocol coded data based on the ICAP theory and teacher instructional approaches. Over the course of six weeks, data on student engagement behaviors and teacher behaviors were gathered. In addition to using coding and counting methods, Epistemic Network Analysis (ENA) was employed to explore the trajectory of the STEM-PBL course during these six weeks. The results revealed that passive student engagement behaviors and teacher- centered behaviors clustered on the left side, while active student engagement behaviors and student-centered teacher behaviors clustered on the right side. Furthermore, examining the network graphs for each week's courses showed that as the course progressed, instruction shifted from teacher-centered to student-centered, and students exhibited higher levels of engagement. These findings provided insights into the dynamics of STEM-PBL classrooms.

With the rise of online learning, multiple-text reading has become a prevailing trend and involves complex cognitive processes. Previous research often employed controversial socio-scientific issues in the form of expository texts to examine reading processes and comprehension. However, the story is relatively accessible, which can enhance learners’ interest in reading. Moreover, the reading order could affect how learners integrate multiple texts. Therefore, in this study, we employed eye- tracking technology to investigate how different genres and reading orders of conflicting texts influence readers' change of interest, learning performance, and eye movement patterns. The study included 21 participants who were divided into two groups: expository text first and story first. Each participant was provided with two stories and two expository texts to read. Their interests were recorded before and after reading, alongside their eye movements were recorded during the reading. After completing the reading, participants were required to write an essay and take a multiple-choice test to assess their comprehension. The findings revealed that, regardless of reading order, participants’ interest heightened after reading. however, the expository-text-first group had higher writing scores than the story-first group. As for the eye movement measures, this study showed that although the first-pass reading time was not affected by reading order and genre, the rereading time was. Specifically, the expository-text- first group spent more rereading time on the expository text compared to the story-first group, indicating that these readers devote more rereading time to the expository text, focusing on taking its basic concepts and theories, resulting in greater integrated comprehension. Therefore, they exhibit superior performance in deep-level reading comprehension. This study demonstrates that reading order influences readers' reading comprehension, providing useful insights for future scientific topic education and discussions.

Complex emotion is an aggregate of two or more others which has highly variable appearances, inter-dependence, and affective dynamics.These properties make the recognition hard to handle via existing recognition techniques like action units or valence-arousal detection. In this study, we propose a bionic two-system structure for complex emotion recognition. The structure mimics the working theory of the human brain responding to problems decision-making. System I is a fast compound sensing module. System II is a slower cognitive decision module that processes data more integratively. System I contains one branch for facial expression feature representation including basic emotion, action units, and valence arousal detection and one for physiological measurement which is an image-only implementation for practicality. In System II, a decision module with segmentation is employed to ensure the chosen period including the emotion occurrence and iteratively optimize the emotion information in a given segment via reinforcement learning. The proposed method outperforms state-of-the-art on emotion recognition tasks with an accuracy of 94.15% in basic emotion recognition on the BP4D and an accuracy of 68.75% for binary valence arousal classification on the DEAP. For a subset of complex emotions, the recognition accuracy exceeds 70% on both databases, that is a significant improvement.

Compared to reading text alone, pictures are regarded as easier for readers to comprehend the context. For EFL readers, their reading behavior on text with pictures needs to be carefully inspected. The study aims to examine how different are the viewing behaviors of EFL beginners versus intermediate readers on reading narrative paragraphs and accompanying pictures. Seventeen junior high and twenty-one senior high students represented as EFL beginners and intermediate readers, respectively. Both of them read consecutively three screens with narrative texts and pictures illustrating the texts. The results showed that both beginners and intermediate readers paid more attention to the texts than the pictures. The beginners almost solely fixated on the texts and few fixations fell on to the pictures while the intermediates had more fixations on both texts and pictures. The eye-movement data in the specific AOIs showed that the intermediates made more references between text and pictures when they encountered difficult words or processed semantic meaning making. The beginners were less efficient in reading, having less fixated time on each screen, and encountered greater difficulties in comprehension than the intermediates. Based on eye-movement data, a personalized strategy to alter display sequence could be provided to support EFL beginners: Before going into narrative reading, a reminding message could be dispatched onscreen guiding them to view standalone pictures and to inspect pictorial components carefully to serve as the macro-reading strategy. The personalization could also be realized by posting cognitive and meta-cognitive level questions during the inspection of picture.

Reinforcement learning theory encourages the use of agents for stimulating and assisting learners in their efforts to develop thinking styles. In this study the authors looked at a similar scenario of human-environmental interaction using Internet-mediated simulations as learning environments. One hundred and forty-nine vocational high schools students participated in this study to see if they can develop adequate thinking style when they learned in a simulation environment with help from agents. The results show that the judicial thinking style was best suited to the system we designed for this project - that is, we observed the greatest amount of development for this particular thinking style. Our results indicate that it is possible to establish and support thinking styles via Internet-mediated simulations.

This study applied NetPeas (networked peer assessment system) in peer assessment activities of graduate level of thesis writing course. Authors wish to discover the students' attitudes toward and desired system requirements of NetPeas in writing course. This study involved 100 electrical engineering graduates from two research-oriented universities at northern Taiwan. Those students were required to attend a weekly thesis writing course and submit eight assignments through four rounds of networked peer assessment, and then authors survey their attitudes and opinions toward NetPeas. In the last, quantitative and qualitative surveys indicated that significantly more students willing to join the networked peer assessment activities via NetPeas in the near future and stand for using NetPeas. Authors also have dug several desirable requirements for NetPeas.

Despite their increasing availability on the Internet, networked peer assessment systems lack feasible automatic dispatching algorithm of student's assignments and ultimately inhibit the effectiveness of peer assessment. Therefore, this study presents a reviewer dispatching algorithm capable of supporting networked peer assessment system in order to automatically dispatch student's assignments to their peer reviewers.