This study enhances the modeling of contributor roles in MOOC discussion forums by analyzing networks of information-giving and information-seeking among users. It characterizes user roles based on connection patterns and thematic areas of information exchange, using a socio-semantic blockmodelling approach. The research finds that social and semantic roles are not strongly interdependent. Additionally, sequential pattern analysis, based on the ICAP framework, reveals a strong relationship between engagement patterns and socio-semantic user roles. The study combines traditional blockmodelling with semantic analysis to provide a richer understanding of participant roles.