Based on the concept of learning stickiness, this study constructed a model of influencing factors among the elements of online learning content, interaction, incentive, satisfaction, and learning stickiness from Comprehensive Learning Theory. Structural equation modeling was used to analyze the interaction and influence effects among the factors. It is found that the content, interaction, and incentive in Comprehensive Learning Theory had a significant positive impact on learning stickiness from the total effect analysis. From the direct effect analysis, the influence of content and interaction on learning stickiness was not substantial, but the influence of incentive and satisfaction on learning stickiness was significant. From the perspective of mediation effect analysis, incentive and satisfaction were critical mediating variables for the influence of content and interaction on learning stickiness. This study put forward suggestions and strategies for online teaching, providing a reference for teachers to carry out online education.
In this paper, we used the platform log data to extract three features (proportion of passive video time, proportion of active video time, and proportion of assignment time) aligning with different learning activities in the Interactive- Constructive-Active-Passive (ICAP) framework, and applied hierarchical clustering to detect student engagement modes. A total of 840 learning rounds were clustered into four categories of engagement: passive (n = 80), active (n = 366),constructive (n = 75) and resting (n = 319). The results showed that there were differences in the performance of the four engagement modes, and three types of learning status were identified based on the sequences of student engagement modes: difficult, balanced and easy. This study indicated that based on the ICAP framework, the online learning platform log data could be used to automatically detect different engagement modes of students, which could provide useful references for online learning analysis and personalized learning.
Based on the related theories and research results of learning behavioral engagement, this study constructed an evaluation framework of learning behavioral engagement in live teaching, which included 24 indicators in three dimensions: compliance with norms, learning participation and social participation. A small-class live English learning for younger students on the ClassIn was taken as a case study program. Five younger students attended this English learning course of 16 lessons totaling 950 minutes. The preset indicators were preliminarily examined based on the teaching records and the recorded course data. Then, experts in the field of educational technology were invited to develop the learning behavioral engagement dimensions and indicator weightings by using the Analytic Hierarchy Process, and to determine the evaluation indicator system for the evaluation of learning behavioral engagement. Finally, based on this framework, the characteristics of learning behavioral engagement of the case course were analyzed, and the influences of students’ individual factors, teaching and environmental factors on learning behavioral engagement in live teaching were investigated.
Using log data of 823 university students collected in two settings: their online learning setting and daily life setting (using campus ID cards for consumption purposes and book-borrowing in the university library), this study created indicators for online learning behavior, early-rising behavior, book-borrowing behavior and learning performance prediction. Five machine learning models were employed to analyze learning performance prediction, with the additional use of Boosting and Bagging to improve the accuracy of the prediction model. The predictability of the proposed model was also compared with that of both the Artificial Neural Network model and the Deep Neural Network model. At the same time, a classification rule set was established by combining decision tree and rule model, and a learning behavior diagnosis model combining decision tree and deep neural network was constructed. Findings showed that multi-scenario behavior performance indicators had strong predictive capabilities while the Deep Neural Network model had the highest prediction accuracy (82%) but was most time-consuming. The model based on the rule set is highly accurate, readable and operable and may be conducive to making accurate teaching interventions and resource recommendations.
Discussion is a common and important learning process. Involvement of a virtual agent can provide adaptive support for the discussion process. Argumentative knowledge construction is beneficial to learners’ acquisition of knowledge,but the effectiveness of argumentative scaffolding in existing studies is not consistent. Based on an intelligent discussion system, a total of 47 undergraduate students took part in the experiment and they were assigned to three different conditions: content-related plus content-independent scaffolding condition, content-related scaffolding condition, and the control condition. Under the content-related and content-independent scaffolding condition, the computer agent provided an idea from semantically different categories (content-related scaffolding) according to the automatic categorization of the current contributions, and further inquired the participants about their attitudes and reasons (content-independent scaffolding). Under the condition of content-related scaffolding condition, the virtual agent only provided semantically different viewpoints. Under the control condition, the subjects expressed their opinion independently without the participation of the virtual agent. Findings revealed that compared with the control group, when the virtual agent provided semantically different ideas (content-related scaffolding), the discussion breadth (number of categories) was improved and the subjects felt that they had a more comprehensive understanding of the problem. Compared with the content-related scaffolding condition, when the virtual agent provided semantically different ideas and further asked about the attitudes and reasons, the subjects expressed more agreement with these views, but mentioned fewer categories during the discussion. This study suggests that the content-related scaffolding can facilitate the cognitive processing effect relevant to the topic of discussion. When the content independent scaffolding is added, it can promote the argumentative processing, but may have a negative effect on the cognitive processing related to the topic discussed.
The eye-tracking technology was used in this study to investigate the effects of embedded questions and feedback in instructional videos on learning performance and attention allocation and whether an expertise reversal effect existed. The experiment involved 49 learners with high-level prior knowledge and 45 ones with low-level prior knowledge from a university. Meanwhile, they learned instructional videos with no embedded feedback, embedded questions without feedback and embedded questions with feedback. Findings from the experiment showed that the instructional videos with embedded questions but without feedback not only improved the participants’ attention but also enhanced their learning performance. Furthermore, there was an expertise reversal effect on the learning performance whereby instructional videos with embedded questions but without feedback improved the learning performance of learners with low-level prior knowledge, but not those with high-level prior knowledge.