Research On Participant Model In Computer Supported Collaborative Learning

Collaborative learning refers to a form of learning in which two or more individuals participate in learning together to achieve learning goals.It is a form of learning in which participants freely share ideas,feelings and attitudes for a specific topic,thereby achieving knowledge,skill growth and attitude transformed forms of mutual learning.Compared with traditional individual learning,collaborative learning has the following advantages:in terms of skills,(1)it can develop higher-level cognitive skills;(2)it can increase the skills of importing expression and communication;additionally,(3)creative thinking can be developed;(4)emotional attitude changes can be achieved;(5)collaborative learning also provides an opportunity to expand the available information pool,so that the group can obtain higher quality than anyone can achieve solution.However,in the process of collaborative learning,not all participants had the same learning effect,and there were differences in knowledge acquisition among participants who play different roles or participants type in collaborative learning.Building a participant model was an important topic in collaborative learning.With the development of the Internet and computer technology,computer-supported collaborative learning(CSCL)had received more and more attention from researchers,and it had also provided opportunities for scientifically and accurately building a participant model in collaborative learning.The present research aims to explore the variables to building participant model in CSCL,which mainly includes the following two aspects:(1)Analyze the behavior of participants in CSCL(used to building the participant model).Firstly,based on the perspective of positive interdependence,the behavior of participants namely ICCC variables(Individual Conversation Characteristics Curves variables:Participation,Response time,Internal cohesion,Overall responsivity,Social impact,and Communication density)in CSCL was analyzed.Secondly,text analysis and social network analysis were used to calculate participants'scores in ICCC variables,and latent profile analysis,potential profile transition analysis,and conformance analysis were used to test the effectiveness of ICCC variables to building participant modeling.(2)Examine the validity of participant modeling related variables(ICCC variables)in CSCL,and predictive variables of ICCC variables.On the one hand,the relationship between individual factors(e.g.,proactive personality,self-efficacy,positive emotion and task value)and ICCC variables was investigated.On the other hand,the relationship between individual-group congruence and ICCC variables was investigated.The present using text analysis,social network analysis,path analysis,latent profile analysis,latent profile transition analysis and polynomial regression to analyze the data,the results show that:(1)ICCC variables can accurately construct the participant model in CSCL;(2)The individual's proactive personality plays a significant positive role in predicting participation,internal cohesion,overall responsivity,communication density and social impact;Task value was positively associated with participation and overall responsivity;Self-efficacy was positively correlated with participation,internal cohesion,overall responsivity,communication density and social impact;Positive emotion plays a significant positive role in predicting individual participation and response time,and negatively associated with overall responsivity;(3)On the line of individuals-group congruence,as the level of proactive personality increases,the individual's participation,response time,and internal cohesion will also increase.As the level of self-efficacy increases,participation,response time,overall responsivity,social impact,and communication density will also increase.As the level of positive emotions increases,participation,response time,and social impact will also increase.As the task value level increases,participation,response rate,internal cohesion,and overall responsivity will also increase.On the incongruence line of individual-group,when the group's proactive personality is higher than the individual's proactive personality,the individual is more involved in collaborative learning;when the individual self-efficacy is incongruence with the group self-efficacy,the higher their level of incongruence,the higher the social impact,and the lower the communication density.When individual positive emotions are higher than group positive emotions,the individual's internal cohesion and overall responsivity are also higher.The results of this study showed that the participants to discuss the text and the social relation network mode characterized its behavior pattern in CSCL,individual factor,group factor and the interaction between them will affect the participants behavior in CSCL,had affected participants modeling in CSCL.More broadly,the present study provides a framework to explore participants'internal cognitive processes,interpersonal patterns,and social cognitive processes,and provides theoretical and empirical evidence for the construction of participant models,as well as future related theoretical research and the construction of "adaptive" collaborative learning systems.