dc.description.abstract | During non-face-to-face online collaboration, it’s hard to be conscious of the information shared within students. They don’t know where their partners are looking at. Thus, free-riding often occurs, and that leads to the deficiency of organization, coordination and high-quality conversation. Therefore, we collected 128 participants between 18 to 25 in groups of two in this study. We used two-way ANOVA with sychronnous control mode (SC) and distributed control mode (DC) data, whether provided attention guiding tool (AGT) or not, to understand how students collaborative deal with the scientific problem by computer simulation. We used eye-tracking to analyze the student’s eye movement to understand their attention distribution and learning behavior. Also, we quantified the discourse to find out their engagement and used questionnaire to collect their feelings. Furthermore, we revealed the relationship between group and student by Hierarchical Linear Modeling to realize how collaboration affects students. In addition, we used Pearson Correlation Coefficient to understand the relationship between group variables. The results show that students′ gaze duration, gaze frequency, feelings and discourse have different effects. Otherwise, providing AGT will also affect the students′ gaze fixation. However, there are different effects in their performance and discourse when AGT is providing. We find that discourse is closely related to personal gaze fixation and feelings. Students exchange information more frequently, expose each other’s attention and clarify the phenomenon easier during activity after we provide AGT. Nevertheless, providing AGT in SC mode and DC mode has different effects. It’s easy to neglect some important information in the SM, while it can ameliorate the complicated collaboration to help students find important information in DM. To sum up, AGT needs to be used with caution. | en_US |