Network Patterns of Direct and Indirect Reciprocity in edX MOOC Forums

Abstract

This paper proposes a set of indicators to capture the social context between regular forum posters in Massive Open Online Courses (MOOCs). Such indicators advance learning analytics research of social relations in online settings, as they enable to compare interactions in different courses. Proposed indicators were derived through exponential random graph modelling (ERGM) to the networks of regular posters in four MOOC forums. Modelling demonstrated that social context can be captured by the network patterns of reciprocity. In some MOOCs, network formation is driven by the higher propensity for direct reciprocity and lower propensity for indirect and triadic-level reciprocity. The social context in a highly moderated course was described by lower propensity for direct reciprocity and higher propensity for indirect and triadic-level reciprocity. We conclude that patterns of direct, indirect, and triadic-level reciprocity reflect variations of behaviour in knowledge exchange on MOOC forums. These three types of patterns can be theorised as dyadic information exchange, social solidarity, and gradual amplification of information flow. Re-modelling the same four MOOC networks without the staff and teaching assistants suggested that these network actors play a role in the formation of indirect and triadic-level reciprocity patterns, related to group cohesion.