Quantifying social presence in online-based learning: a statistical and didactical analysis of indicators from social network analysis

Abstract

Social presence is one key factor for successful learning in socio-constructivist learning theories. Teachers need easily interpretable and didactically relevant information to monitor social presence and to intervene if required while a course is running. Social network analysis is a promising method to provide this information, yet it is unclear which indicators are most appropriate for this aim. This study aimed to identify and evaluate indicators derived from social network analysis, which could quantify social presence and provide didactically relevant information to teachers. 3,546 postings from different modules and study groups (n=49) of an online-based Master course were manually coded. Egocentric measures were calculated from social network analysis. Path models were developed to analyze which indicators from social network analysis were statistically significant measures of social presence. The course of these indicators was analyzed across the modules and statistically evaluated using the Friedman test. Out of 13 possible indicators, six indicators were found to be statistically and didactically relevant (Ties, Density, Efficiency, nBroke, Out- & In-closeness). These indicators showed high regression weights in the path models and the development across multiple time points was identified to be statistically significant (p < .001). Literature analysis and didactical considerations showed that these indicators might provide an appropriate and real-time overview of students' social presence. We identified indicators that allow to measure social presence and provide didactically relevant information for teachers in real time by using log data from learning management systems. Further research will now focus on implementing the findings in a teacher dashboard.